Change Log

v5.0.0-beta.14

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.14
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.14
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.14

Contact us to enable all Enterprise Features [EE]

---

• Release July 17th 2024

• Support for SVG image templates (backend + Studio).
  Update SVG by modifying specific attributes based on the provided data with these new formatters:

  • :svgUpdate(attrName, selectorValue, selectorType) Updates an SVG attribute (e.g., fill, stroke) of the selected SVG nodes.
  • :svgSelectiveUpdate(attrName, attrValue, selectorType) Selects SVG nodes and injects a value. This formatter works like the previous one but uses a selector as input instead of a direct attribute value.
    

  Parameters:

  • attrName: The SVG attribute to update. This can be:
    • A fixed string in quotes (e.g., fill, stroke).
    • A relative path to a property of the parent object (e.g., .siblingObj.id).
  • attrValue: The attribute value(s) to insert into the SVG. (e.g., a color value like '#F00').
  • selectorValue: The value used to select the target SVG element(s). This can be:
    • A fixed string (e.g., 10).
    • A data path (e.g., .siblingObj.id).
  • selectorType: The SVG attribute used to identify and select shape(s) to modify. Default: id *. Options:
    • id : selects only the node with the exact id.
    • id > * : selects the node and its direct children.
    • id * : selects the node and all its descendants.
      

  Template example:

    <!-- {d.myTable[].width:svgUpdate('width', .id)} -->
    <!-- {d.myTable[active=true].id:svgSelectiveUpdate('fill', .color)} -->
    <svg width="100%" height="100%" viewBox="0 0 596 842" version="1.1">
      <circle id="B12" cx="220" cy="434" r="149" style="fill:#ebebeb;"/>
      <circle id="B10" cx="379" cy="232" r="122" fill="#ababab"/>
      <circle id="B11" cx="406" cy="638" r="95" fill="#ebebeb"/>
    </svg>

  After data injection:

    <svg width="100%" height="100%" viewBox="0 0 596 842" version="1.1">
      <circle id="B12" cx="220" cy="434" r="149" style="fill:#ebebeb;" width="200"/>
      <circle id="B10" cx="379" cy="232" r="122" fill="#66a200" width="100"/>
      <circle id="B11" cx="406" cy="638" r="95" fill="#aa66aa" width="300"/>
    </svg>

  Data sample

      {
        "id"      : "B10",
        "myTable" : [
          { "id" : "B10", "color" : "#66a200", "active" : true  , "width" : 100},
          { "id" : "B12", "color" : "#aa66aa", "active" : false , "width" : 200},
          { "id" : "B11", "color" : "#aa66aa", "active" : true  , "width" : 300}
        ]
      }

  Only basic SVG shapes are updated: rect, circle, line, ellipse, polygon, polyline, and path.

---

• 🚨 Major Release 🚨

  All changes described below mainly concern developers who integrate Carbone Studio into their own solutions.
  Some of these changes may evolve based on feedback from early users. For all other users, everything will continue to work as before (plugins, SDKs, API, document generation, etc.).

• [Studio] New Studio starting modes: <carbone-studio carbone-mode="embedded">

  • embedded (default): Displays only the designer page, without template version management. Intended for integrators embedding the Studio.
  • light: Same as embedded, but with the Carbone title bar. This is the default Studio interface when opened in a browser without any database configured.
  • full: Displays all pages and full navigation. Available only if a database is defined (see below). This mode is the complete interface that will replace the Studio in the Cloud Edition.

• [Studio] Add the ability to override fetch requests made by the Studio

    const studio = document.createElement('carbone-studio');
    window.addEventListener('DOMContentLoaded', function() {
      document.body.appendChild(studio);
      studio.setConfig({
        fetchOverride : fetchHandler // Override fetch behavior in the Studio 
      });
      function fetchHandler(url, options) {
        console.log('Override fetch called:', url, options);
        const newOptions = {
          ...options,
          headers: {
            ...options.headers,
            'X-From-Studio': 'true'
          }
        };
        if (url.startsWith('/template/')) {}
        if (url.startsWith('/templates/')) {}
        if (url.startsWith('/render/')) {}
        return fetch(url, newOptions);
      }
    });

• [Studio] New live reload behavior and new events for integrators embedding the Studio

  When live reload is active and the user is editing a template, the Studio now sends ephemeral templates (technically, only POST /render is called, no POST /template).
  When live reload is turned off, the user is prompted to save their changes.
  They can view a list of their changes and switch between the previous and current template versions.
  For integrators, a new event can be listened to in order to detect when the user clicks the Save button:
  

    // Listen for the `template:saved` event to get the new template Id
    studio.addEventListener('template:saved', handler);
    // Event object:
    // {
    //   detail: {
    //     templateId: 'af42af42af42af42af42af42af42af42af42af42af42af42'
    //     type      : 'docx'
    //   }
    // }
  
    // The `template:updated` event is still emitted, but it contains the template as a Base64 string (no template ID)
    studio.addEventListener('template:updated', handler);
    // Event object:
    // {
    //   detail: {
    //     dataURI : '',   // template in base64
    //     type    : 'docx'
    //   }
    // }

• [Studio] Customize Studio CSS theme

  The component accepts a custom attribute where CSS variables can be passed to override default colors.

  • Method 1: JavaScript

      studio.setAttribute('carbone-theme', 'main { --primary: #000; }');

  • Method 2: HTML

      <carbone-studio carbone-theme="main { --primary: #000; }"></carbone-studio>

  Here is a quick overview of all available CSS variables:

    /* Beer CSS main colors - Material Design */ 
    --primary:#a544c5;
    --on-primary:#ffffff;
    --primary-container:#eecafc; 
    --on-primary-container:#350040;
    --secondary:#6b586b;
    --on-secondary:#ffffff;
    --secondary-container:#cac5ca;
    --on-secondary-container:#251626;
    --tertiary:#9e9e9e;
    --on-tertiary:#ffffff;
    --tertiary-container:#ffdad2;
    --on-tertiary-container:#32110c;
    --error:#ba1b1b;
    --error-container:#ffdad4;
    --on-error:#ffffff;
    --on-error-container:#410001;
    --background:#fcfcfc;
    --on-background:#1e1a1d;
    --surface:#f5f5f5;  
    --on-surface:#1e1a1d;
    --surface-variant:#e8e5e9;
    --on-surface-variant:#1e1a1d;
    --outline:#a7a7a7;
    --inverse-surface:#323639;
    --inverse-on-surface:#f7eef3;
    --inverse-primary:#fbaaff;
    --inverse-on-primary: #371E73;
    --outline-variant: var(--outline);
    --surface-dim: #f4f4f4;
    --surface-bright: #fdf8fd;
    --surface-container-lowest: #ffffff;
    --surface-container-low: #f7f2f7;
    --surface-container: #fff;
    --surface-container-high: #ece7eb;
    --surface-container-highest: #e6e1e6;
    --active: rgba(0,0,0,.1);
    --overlay: rgba(0,0,0,.5);
    /* Studio tab colors */
    --tab-preview-border-color: #999;
    --tab-border-color: #c0aabf;
    --tab-text-color: #999;
    /* JSON Editor colors */
    --json-editor-active-line: rgb(235 204 255 / 27%);
    --json-editor-active-bracket: rgb(193 96 255 / 27%);
    --json-editor-search-match-selected: rgb(237 178 255 / 51%);
    --json-editor-search-match: rgb(255 211 0 / 35%);

• [Studio] New JavaScript API for interacting with the embedded Studio

  Get or set the rendering options in the Studio using the studio.setRenderOptions(object) and studio.getRenderOptions() functions.
  Calling setRenderOptions() updates the user interface but does not trigger rendering events. A message will indicate to the user that the preview must be refreshed manually. Example:

    studio.setRenderOptions({
      data           : {},             // Object or Array
      complement     : {},             // Object or Array
      enum           : {},             // Object
      translations   : {},             // Object
      lang           : 'en-us',        // Lowercase Language Culture Code
      timezone       : 'Europe/Paris', // Timezone Identifier
      currencySource : 'EUR',          // Currency CODE
      currencyTarget : 'EUR'           // Currency CODE
    })

• [Studio] Miscellaneous User Experience Improvements

  • Improved drag-and-drop design: you can now also click the drop zone to upload a file
  • Added support for dragging and dropping JSON directly into the JSON panel (automatically replaces the previously loaded JSON)
  • Improved icon weight and colors
  • Fixed a bug that caused double extensions when downloading the generated report
  • Fixed various other bugs
  • UX reworked with two clear actions, "New" button to create a new template, "Edit" button to update an existing template
  • Warns users before leaving the Studio without saving their current work
  • Fixed white page issue by respecting the Accept-Encoding header. Now supports Brotli, Gzip, and uncompressed requests

• [Studio] New HTML Editor + HTML Viewer

  Modify HTML templates directly from the Studio instead of using an external tool, with live reload. The Studio includes a lightweight WYSIWYG HTML editor. Ideal for playground use or quick edits to existing HTML templates. More features (such as image support) will be added in the future.

• [Studio] New XML/TXT Viewer

  A new viewer for previewing generated TXT and XML files directly in the Studio. The Studio now supports 3 viewers: HTML, PDF, and XML/TXT. It automatically selects the best viewer based on your template.

• ⚡️ [Studio+Backend] Template management embedded in the On-Premise solution

  A new parameter databaseName: "database.sqlite" can be defined in the configuration (or using the environment variable CARBONE_DATABASE_NAME).

  If this database is set, it enables all features related to template management, including:

  • Listing templates
  • Storing metadata with templates
  • Using both templateId (constant id) and versionId
  • The full version of the studio

  In this version, the database is stored in the config/ directory.
  This will change in the final release: the database will be stored like a template and automatically synchronized across multiple nodes if needed.

• [Backend] New API for managing templates

  👋 Available only for:

  • On-Premise version (Cloud support coming soon)
  • When "stateful" mode is enabled - technically, if a database name is defined in the configuration
  • Experimental use only: all existing data may be deleted in the next stable release

  What problem does it solve?

  • Allows keeping a "constant" template ID
  • Lets you decide when a new version of a template is deployed to production
  • Enables listing and managing all templates via the API
  • Allows storing optional metadata (name, comment, etc.) and sample data alongside templates
  • Unifies the user interface (Studio) between the On-Premise and Cloud versions

  What's new?

  All API endpoints now support two types of IDs when targeting a template:

  • Version ID
    A SHA-256 hash of the template file, automatically generated by Carbone.
    This is the default behavior in Carbone v2, v3, and v4.
    It matches the filename of the template stored on disk (or S3).
    Formerly called templateId in our documentation.

  • Template ID A short and constant 64-bit ID, automatically generated by Carbone, and optionally defined by the user.
    Multiple template versions can share the same template ID.
    When calling endpoints with a template ID (e.g., POST /render/{ID}), Carbone will use the latest deployed version (based on the deployedAt timestamp). The previous behavior is still available: you can call POST /render with the versionId to target a specific version.

  To summarize:

  To maintain backward compatibility, this new method for storing templates is only enabled if the boolean versioning: true is set when creating a template via POST /template. In that case, the response of POST /template is updated to avoid confusion:

  Type	No Versioning	Versioning Enabled	Carbone Version
  SHA-256 hash of the file	templateId	–	v5, v4, v3, v2
  SHA-256 hash + random salt	–	versionId	v5 + versioning: true
  64-bit constant ID	–	id	v5 + versioning: true

  API updates summary

    # Updated endpoints:
    POST   /template                            # Create a new template or a new version of an existing template
    GET    /template/{templateId-or-versionId}  # Download the template file
    DELETE /template/{templateId-or-versionId}  # Delete a template
    POST   /render/{templateId-or-versionId}    # Generate reports using templateId or versionId
  
    # New endpoints:
    PATCH  /template/{templateId-or-versionId}  # Update metadata: name, comment, tags, expireAt, deployedAt
    GET    /templates                           # List all deployed templates and their versions
    GET    /templates/categories                # List all categories used in templates
    GET    /templates/tags                      # List all tags used in templates

  POST /template

  • Adds the ability to store metadata along with the template.
  • Allows attaching multiple template versions to the same templateId.
  • Lets you specify which version is currently deployed using deployedAt. When a report is generated using the new template ID, Carbone selects the template version with the highest deployedAt timestamp that is not in the future.
  • For backward compatibility, the new parameter versioning: true must be set in the request body to enable these features.

  Request body:

  {
    // Required
    template   : base64 or multipart/form-data,
  
    // Optional
    versioning : true,   // Enables new versioning mode. Disabled if undefined or false, and if "id" is undefined
    id         : "4242", // Template ID. Generated if not provided and versioning = true ()
    name       : "",     // Template name
    comment    : "",     // Template comment
    tags       : [],     // List of tags
    samples    : [],     // Sample input data
    expireAt   : 0,      // UTC Unix timestamp; values ≥ 42000000000 (year 3300) are interpreted as "NOW". 0 means it never expires.
    deployedAt : 0       // UTC Unix timestamp; values ≥ 42000000000 (year 3300) are interpreted as "NOW". Value in the future are not allowed.
  }

  Reponse body

  The new response when versioning = true is used

  {
    success    : true,
    data       : {
      id         : "4242"
      versionId  : "af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42",
      type       : "docx",
      size       : 12012,
      createdAt  : 21545451
    }
  }

  For backward compatibility, when versioning is set to false or is undefined,
  Carbone returns only the SHA-256 hash of the template file like in Carbone v4.

  {
    success    : true,
    data       : {
      templateId : "af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42"
    }
  }

  PATCH /template/{templateId-or-versionId}

  • Same documentation as POST /template, except:
  • Only the metadata fields name, comment, tags, expireAt, and deployedAt can be modified. All other fields are ignored.

  GET /template/{templateId-or-versionId}

  Accepts both the new templateId and the versionId (as in Carbone v4).

  • GET /template/{templateId} downloads the latest deployed template file (64-bit number).
  • GET /template/{versionId} downloads a specific version of the template (SHA-256 file hash).
  • GET /template/{templateId}_sample.json downloads the data sample stored with this template (used in the Studio).
  • GET /template/{versionId}_sample.json downloads the data sample stored with this specific version (used in the Studio).

  DELETE /template/{templateId-or-versionId}

  Accepts both the new templateId and the versionId (as in Carbone v4).

  • DELETE /template/{templateId} deletes the latest deployed template and all its previous versions (64-bit number).
  • DELETE /template/{versionId} deletes only the specified version (SHA-256 file hash). Technically, Carbone updates the expireAt field when this endpoint is called and garbage-collects the template after a delay (default: 24 hours). This has the same effect as calling PATCH /template/{id} with expireAt = NOW.

  GET /templates

  List all templates. By default, it returns all deployed template only, sorted by createdAt DESC, with expireAt > now

  Reponse body

  {
    success    : true,
    hasMore    : false, // Whether or not there are more elements available after this set. If false, this set comprises the end of the list.
    nextCursor : "",
    data    : [
      {
        id         : "4242",
        versionId  : "af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42af42",
        createdAt  : 1751385172,
        expireAt   : 0,
        deployedAt : 1751385172,
        size       : 12012,
        type       : "docx",
        name       : "template name",
        comment    : "comment"
        category   : "clientA"
        tags       : ["prod", "other"]  
      }
    ]
  }

  Query parameters

  • ?id=number filter by template id
  • ?versionId=number filter by version Id
  • ?category=string filter by category
  • ?includeVersions=boolean By default equals to false, Use true to list all versions for a specific template Id
  • ?search=string search in template name (fuzzy search), version Id (exact) or template id (exact)
  • ?limit=number limit the number of items (100 by default)
  • ?cursor=string A cursor to use in pagination. nextCursor is an object ID that defines your place in the list. For example, if you make a list request and receive 100 objects with nextCursor=abcdef, your subsequent call can include cursor=abcdef to fetch the next page of the list.

v5.0.0-beta.13

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.13
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.13
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.13

Contact us to enable all Enterprise Features [EE]

---

• Release: June 30th, 2025
• [EE] New features and improvements for the :html formatter available for ODT/DOCX/PDF:
  • CSS Style Attributes: Added support for rendering inline CSS style attributes color and background-color, even when the HTML is in inline mode, for example: <p style="color:green;background-color:#ff0000">OK</p>. Styles defined within a <style> element, or those applied through classes and IDs, are not supported. If a color is already applied to the template paragraph, the CSS color style will override it. Supported color formats include:
    • Hexa: The hexadecimal color syntax of an sRGB color. Example: #FF00BB.
    • HSL: Expresses a color in the sRGB color space according to its hue, saturation, and lightness components. Examples: hsl(120deg 75% 25%), hsl(120 75 25) or hsl(120,75,25).
    • RGB: Expresses a color in the sRGB color space according to its red, green, and blue components. Examples: rgb(31 120 50) or rgb(31,120,50);.
    • Named Color: The name of a color. Example: red, blue, black, or lightseagreen.
  • Support for Optional End Tags: Optional end tags "p" and "li" are now supported. An example structure: <ol><li><p>text<li><p>text2</ol>.
  • Support Tab Entities: Include the character entities   or   to create a tab in an ODT/DOCX/PDF. Do not use 	, as HTML parsers will collapse it into a single space due to the whitespace collapse principle.
  • Bug Fix: Fixed an issue with nested ordered/unordered list items mixed with empty paragraphs.

v5.0.0-beta.12

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.12
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.12
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.12

Contact us to enable all Enterprise Features [EE]

---

• Release: June 27th, 2025
• [EE] Improved support for filling PDF forms:
  • Ignore errors when the PDF form contains buttons or undefined text fields
  • Accept /Text and /FreeText annotations
  • Remove annotations that contain Carbone tags with the :checkField and :fillField formatters
    (previously, only tags with :check and :fill formatters were removed)

v5.0.0-beta.11

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.11
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.11
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.11

Contact us to enable all Enterprise Features [EE]

---

• Release June 17th 2025

• [EE] Resolved an issue with the :html formatter when ol/ul lists are mixed with br tags.

• [EE] 📄 PDF templates are supported to fill forms with your data. This is a low-level feature to update PDF forms. A more user-friendly solution will be available in the final version of the Studio. Three methods are available:

  • Place Carbone tags directly in the form fields
    Insert Carbone tags into the text fields of your PDF form. Carbone replaces them with data from your dataset. Loops are allowed in text fields.

  • Add annotations above the fields you want to update, using formatters like :fill, :check, or :uncheck. Carbone will:

    • Detect the form field below the annotation,
    • Remove the annotation,
    • Update the corresponding field. If the field is not found, an error is returned. In that case, move your annotation to make sure it overlaps the target field.

    Examples:

    • {d.myText:fill} fills the text field behind the tag.
    • {d.myCondition:ifEQ(true):check} checks the radio button or checkbox behind the tag.

  • Directly target form fields by name using formatters
    Place Carbone tags anywhere in a PDF annotation using formatters like :fillField('fieldName') or :checkField('fieldName'). Examples:

    • {d.text:fillField('fieldName')} fills the text field with the provided value. It can also check a checkbox (any non-empty value will check it).
    • {d.genre:ifEQ('boy'):show(male):ifEQ('girl'):show(female):fillField('fieldName')} selects the corresponding option in a radio group ('male' or 'female'). If the option does not exist, an error is returned.
    • {d.text:ifEQ(true):checkField('fieldName')} checks the checkbox if the condition is true.

  • Known limitations:

    • In the Studio, the embedded PDF viewer cannot display filled forms. To preview results, uncheck the "Use embedded PDF viewer" option in the preview panel.
    • Currently, Carbone supports only Text fields, Radio buttons, and Checkboxes.
    • Warning: The native macOS PDF viewer can break PDF forms that include radio buttons.

• ⚡️ Define custom HTML templates for PDF headers and footers, specifically for generated PDF with the Chrome converter (converter: 'C'). Templates must include the custom HTML elements <carbone-pdf-header> and <carbone-pdf-footer>, for example:

    <carbone-pdf-header>
      <div style="font-size:12px; font-family:Arial; width:100%; text-align:center;">
        My Header text
      </div>
    </carbone-pdf-header>
  
    <carbone-pdf-footer>
      <div style="font-size:12px; font-family:Arial; width:100%; text-align:center;">
        My footer text
      </div>
    </carbone-pdf-footer>

  • Special tags can be used to show the current page number and the total number of pages:
    • Current page number: <span class="pageNumber"></span>
    • Total number of pages: <span class="totalPages"></span>
  • These sections can include Carbone tags, allowing your headers and footers to display dynamic content.
  • They are processed correctly even if they are:
    • Inside HTML comments (<!-- -->)
    • Hidden using CSS (display: none)
  • CSS styles must be inlined and placed directly inside the header and footer sections.

• 👌 Change the output page format when converting from HTML to PDF using the Chrome converter (converter: 'C'). Add a custom HTML element called <carbone-pdf-options> in your template:

    <carbone-pdf-options  paper-size="A3"  landscape="true" margin-bottom="2.5" />

  • This section can include Carbone tags

  • They are processed correctly even if they are:

    • Inside HTML comments (<!-- -->)
    • Hidden using CSS (display: none)

  • Below is a list of all available options:

    Option	Type	Description
    landscape	boolean	Paper orientation. Defaults to false.
    display-header-footer	boolean	Display header and footer. Defaults to false, or true if header/footer is provided
    print-background	boolean	Print background graphics. Defaults to true.
    scale	number	Scale of the webpage rendering. Defaults to 1.
    paper-size	string	Letter, Legal, Tabloid, Ledger, A0, A1, A2, A3, A4, A5, A6
    paper-width	number	Overwrite Paper width in inches. Defaults to 8.27 inches (A4)
    paper-height	number	Overwrite Paper height in inches. Defaults to 11.7 inches (A4)
    margin-top	number	Overwrite Top margin in inches. Defaults to 0.5 inches.
    margin-bottom	number	Overwrite Bottom margin in inches. Defaults to 0.5 inches.
    margin-left	number	Overwrite Left margin in inches. Defaults to 0.5 inches.
    margin-right	number	Overwrite Right margin in inches. Defaults to 0.5 inches.
    page-ranges	string	Paper ranges to print, e.g., 1-5, 8, 11-13. Pages are printed in document order. No page
    		appears more than once. Empty string prints entire document. Invalid ranges are ignored.
    		An error is thrown if start > end.
    prefer-css-page-size	boolean	Prefer CSS-defined page size. Defaults to false,
    		in which case the content will be scaled to fit the paper size

v5.0.0-beta.10

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.10
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.10
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.10

Contact us to enable all Enterprise Features [EE]

---

• Release May 16th 2025
• Includes changes from v4.25.9

v4.25.9

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.9
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.9
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.9

Contact us to enable all Enterprise Features [EE]

---

• Release May 16th 2025
• Fixed batch processing for On-Premise environments
• Added support for base64 data-URI in appendFile and attachFile

v5.0.0-beta.9

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.9
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.9
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.9

Contact us to enable all Enterprise Features [EE]

---

• Release April 24th 2025
• Fix: The :appendFile(position) formatter now inserts pages in the correct order, as in v4.25.8

v4.25.8

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.8
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.8
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.8

Contact us to enable all Enterprise Features [EE]

---

• Release April 24th 2025
• The formatter :appendFile(position) now accepts a position parameter that determines where the external document is inserted in the generated output. Valid values for position are "start" (to insert at the beginning) and "end" (default, inserts at the end).

v5.0.0-beta.8

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.8
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.8
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.8

Contact us to enable all Enterprise Features [EE]

---

• Release: April 22nd, 2025
• The formatter :appendFile(position) now accepts a position parameter that determines where the external document is inserted in the generated output. Valid values for position are "start" (to insert at the beginning) and "end" (default, inserts at the end).
• Improved stability of signature position detection
• Detecting multiple signatures in large documents is now about 100 times faster

v5.0.0-beta.7

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.7
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.7
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.7

Contact us to enable all Enterprise Features [EE]

---

• Release April 14th 2025
• Includes v4.25.7
• Print logs when Chrome fails to start.
• Avoid generating a public/private key pair when a public key is already provided via the CARBONE_AUTHENTICATION_PUBLIC_KEY or CARBONE_EE_AUTHENTICATION_PUBLIC_KEY environment variable

v4.25.7

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.7
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.7
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.7

Contact us to enable all Enterprise Features [EE]

---

• Release April 14th 2025
• Fix: Prevent a crash when a barcode is empty and the svg: true option is used.
• Fix: LibreOffice changed the default value of ExportFormFields. It used to be true, but in the latest version, it is now false. This release restores the previous default value to avoid breaking changes when generating PDFs with forms.

v5.0.0-beta.6

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.6
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.6
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.6

Contact us to enable all Enterprise Features [EE]

---

• Release March 20th 2025
• Improved support for HTML5 templates. Side effect: Carbone adds a trailing slash to void elements (e.g., <br> becomes <br/>).
• Studio updates:
  • Show metrics in logs: time spent on merging, converting, and downloading images/files, as well as the size of downloaded images/files.
  • Added an option to switch between all converters and an option to display the preview as HTML instead of PDF.
• ⚡️ New Document converters: Chrome for more robust HTML to PDF conversion, along with OnlyOffice and LibreOffice (default) for all other documents.
  Now, you can choose from three different document converters, all managed directly by Carbone.
  A new converter option is added in v5 to select the document converter in the studio or directly from the API:

  {
    "converter": "L", // L: LibreOffice (default), O: OnlyOffice, C: Chrome
    "convertTo": "pdf",
    "data": {}   
  }

  Historically, when Carbone started, it automatically tried to detect a LibreOffice binary on the system for document conversion. Three new parameters have been added to control this behavior for LibreOffice, OnlyOffice, and Chrome. By defaults, Carbone starts with these values:

    libreOfficePath : 'auto', // Auto-detects LibreOffice. Leave empty to disable, or set "sofficeExecPath, pythonExecPath".
    onlyOfficePath  : '',     // Disabled. Use "auto" or set "x2tPath, AllFontsPath, fontPath".
    chromePath      : '',     // Disabled. Use "auto" or set "pathToChrome".

  These parameters can also be set using the corresponding environment variables: CARBONE_EE_LIBRE_OFFICE_PATH, CARBONE_EE_ONLY_OFFICE_PATH, CARBONE_EE_CHROME_PATH More details will be available in the documentation. Carbone automatically manages multiple threads for each converter, ensures that all security barriers are active, manages timeouts, and more. Current Limitations for Chrome:
  • JavaScript execution is disabled.
  • Only A4 format output is supported (more formats, with custom margins, headers, and footers, will be available in future releases).
  • All outgoing traffic is limited to GET requests for documents, stylesheets, images, media, and fonts.

v5.0.0-beta.5

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.5
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.5
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.5

Contact us to enable all Enterprise Features [EE]

---

• Release March 11th 2025
• Includes v4.25.6
• When the :sign formatter is used, the signatures array is accessible at the root of the object in the body response of POST /render.

  {
    "renderId" : "random-string.pdf",
    "signatures": [
      {
        "data" : "Bob",
        "page" : 1,
        "x"    : 72.1,
        "y"    : 693.436
      }
    ]
  }

v5.0.0-beta.4

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.4
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.4
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.4

Contact us to enable all Enterprise Features [EE]

---

• Release February 18th 2025
• [EE] Improve the position values returned by the :sign formatter:
  • Return integers instead of floats. Since a point is 1/72 inch, integer precision is sufficient.
  • Set the origin (x = 0, y = 0) at the top-left corner of the page to simplify integration with third-party solutions.
  • Define the x-axis as extending to the right and the y-axis as extending downward.
  • Enhance the precision of the X/Y position by ensuring the returned position corresponds to the top-left corner of the tag.
• [EE] For DOCX and ODT templates, two new HTML formatter options are available:
  • inline: Render HTML within the current DOCX paragraph, at the same location, without moving or adding paragraphs. This option will only render the following HTML tags: a, b, strong, em, i, s, del, u. Other HTML tags are skipped and not rendered. Example usage: {d.value:html(inline)}
  • nospace: Render HTML without adding extra DOCX paragraphs at the end of each <p> tags.
• [EE] For DOCX templates, the HTML formatter has been updated with the following improvements:
  • Hyperlinks (a a tags) are now rendered in headers and footers.
  • Images (a img tags) are now rendered in headers and footers.
  • If the HTML contains ordered (ol) or unordered (ul) lists, there is no need to pre-create a list in the template using your text editor. Carbone will render the list directly in the DOCX format without any manual intervention.
  • When injecting multiple HTML lists using the HTML formatter, numbered lists now maintain consistent ordering. Previously, numbering might continue across multiple different lists.
• [EE] For ODT templates, the HTML formatter is now supported in headers and footers.

v5.0.0-beta.3

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.3
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.3
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.3

Contact us to enable all Enterprise Features [EE]

---

• Release February 3rd 2025
• The new templating engine preserves line breaks in templates! Now, you can create CSV or TXT templates. It fixes line breaks in XLSX files.
• [EE] Added the batchOutput option to the POST /render endpoint, to allow returning a single assembled PDF file instead of a ZIP file during batch processing. Accepted values: pdf or zip (default).
• [EE] New formatter :sign. This can be used with PDF generation to place signatures at specific positions in a document. The POST /render endpoint returns additional details indicating where each signature should be placed in the final PDF, along with its associated data. This information can be shared with any third-party tool to manage the signature process. For example, if the template contains the tags {d.owner:sign} and {d.buyer:sign}, the API returns the following information:

  {
    "debug": {
      "signatures": [
        {
          "data" : "Bob", // Content of {d.owner}. Can be a string, number, object, or array
          "page" : 1,
          "x"    : 72.1,
          "y"    : 693.436
        },
        {
          "data" : "Alice", // Content of {d.buyer}
          "page" : 1,
          "x"    : 72.1,
          "y"    : 693.436
        }
      ]
    }
  }

  In a standard PDF coordinate system, the origin (x = 0, y = 0) is at the bottom-left corner of the page. The x-axis extends to the right, and the y-axis extends upward. The API returns an error, and the report is not generated in the following cases:
  • If the number of injected signatures exceeds 999.
  • If a signature position cannot be determined. The :sign formatter prints a text in the final document in the format "ʕ010". Our recommendation is to set the text color to white in the template to make it invisible in the generated PDF.

v5.0.0-beta.2

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.2
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.2
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.2

Contact us to enable all Enterprise Features [EE]

---

• Release January 30th 2025
• 👋 This is a minor release. A major update is coming soon with exciting features such as Signatures, PDF Merging, a Chrome converter, and Markdown/HTML templates in just a few days.
• Fixed a crash that occurred during a graceful exit when the systemd watchdog was activated.
• Includes all fixes from v4 up to v4.25.3.
• Show an error message if the studio is not running in a secure context.
• Improved Studio:
  • Added text search functionality with a search button and shortcuts:
    • Ctrl+F (open search)
    • Ctrl+G (next occurrence)
    • Shift+Ctrl+G (previous occurrence)
  • Fixed inconsistent icon widths.
  • Added functionality to switch between data/complement/enum/translations.
  • Replaced the save button with a refresh button.
  • Scroll position in the JSON editor/viewer is now preserved.
  • Optimized JSON parsing between JSON editors for better performance.
  • Modifications are now preserved when switching between JSON editor and viewer modes.
  • Fixed language icon display issues on Windows.
  • Prevent leaving the JSON editor when an error occurs.
  • Updated PDF.js from version 4.7.76 to 4.10.38.
  • Reduced gaps between icons for better alignment.
  • Improved the design of the page input field in preview mode.

v5.0.0-beta.1

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.1
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.1
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.1

Contact us to enable all Enterprise Features [EE]

---

• Release December 17th 2024
• Includes v4.25.0

v5.0.0-beta.0

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: available soon

Docker Getting started

• Ready to use: carbone/carbone-ee:full-5.0.0-beta.0
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-5.0.0-beta.0
• Only Carbone (without document converters) : carbone/carbone-ee:slim-5.0.0-beta.0

Contact us to enable all Enterprise Features [EE]

---

• Release October 29th 2024
• Check out the full release notes is here

v4.25.6

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.6
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.6
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.6

Contact us to enable all Enterprise Features [EE]

---

• Release March 11th 2025
• It is no longer necessary to repeat array filters on all tags to filter a loop. Now, if the filter is defined only on the i+1 tag, it is applied globally to the entire loop. This change is activated only when the prerelease tag {o.preReleaseFeatureIn=4024000} is set (by default in v5). To maintain backward compatibility, if at least one filter is defined on the i-th part, the previous rule still applies.
  • Before:

    {d.myArray[i, i>0].id}
    {d.myArray[i, i>0].name}
    {d.myArray[i+1, i>0]}

  • After:

    {d.myArray[i].id}
    {d.myArray[i].name}
    {d.myArray[i+1, i>0]}

v4.25.5

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.5
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.5
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.5

Contact us to enable all Enterprise Features [EE]

---

• Release February 8th 2025
• Fix: Carbone did not escape some forbidden characters in XML when an array was printed in the template without using the arrayJoin formatter. This fix is activated only when the prerelease tag {o.preReleaseFeatureIn=4024000} is set.

v4.25.4

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.4
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.4
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.4

Contact us to enable all Enterprise Features [EE]

---

• Release February 6th 2025
• Fix: Carbone tags in sheet names were not updated since v4.25.3

v4.25.3

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.3
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.3
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.3

Contact us to enable all Enterprise Features [EE]

---

• Release January 28th 2025
• Fixed an issue with corrupted XLSX files caused by defined range names.
• Accepts array filters with colons and &, such as {d.myArray[filter='string&with:colons'].id}, when {o.preReleaseFeatureIn=4022011} is activated. Note: Array filters with commas are still not supported in v4. This will be addressed in v5.
• Formatters containing '&' do not print '&' in the generated document, when {o.preReleaseFeatureIn=4022011} is activated.

v4.25.2

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.2
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.2
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.2

Contact us to enable all Enterprise Features [EE]

---

• Release January 20th 2025
• [EE] Fixed some issues with the :set formatter when {o.preReleaseFeatureIn=4022011} is activated, especially in combination with array filters and conditions. The fix resolves the following cases for :set:
  • {d.myArray[filter=2].val:set(d.result)}: Stores the first value val found in the array that matches the array filter filter=2 (Now the behavior is the same with or without :set).
  • {d.myArray[filter=2].val:ifEM:show(10):set(d.result)}: Stores the first value val found that matches the filter filter=2, otherwise stores "10".
  • {d.myArray[filter=2].val:ifEM:show(10):set(.newAttribute)}: Adds a new attribute in myArray (copy of .val), setting its value to "10" for all items in myArray that do not match the filter filter=2.
• The :convCRLF option is now supported in ODP templates.
• [EE] For ODT/ODS templates, you can now create dynamic bookmarks. To add a dynamic bookmark to a template, right-click on the selected text, navigate to the 'Insert' menu, select 'Bookmark,' and then write the Carbone tag. The value can be a single tag {d.bookmark} or a loop expression {d.list[i].bookmark}. Please note that the following characters are not allowed in a bookmark: /@*?",#.
• Fixed invalid metrics for fetchImageTime, fetchFileTime, fetchImageBytes

v4.25.1

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.1
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.1
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.1

Contact us to enable all Enterprise Features [EE]

---

• Release December 19th 2024
• [EE] Fixed a random crash occurring when multiple documents fail to download while using :appendFile.
• [EE] Fixed crash when using an invalid image URL in injected HTML

v4.25.0

Show download links

Binaries Getting started

• Linux: x64, arm64
• Mac: x64
• Windows: x64

Docker Getting started

• Ready to use: carbone/carbone-ee:full-4.25.0
• Includes 2GB of fonts : carbone/carbone-ee:full-fonts-4.25.0
• Only Carbone (without document converters) : carbone/carbone-ee:slim-4.25.0

Contact us to enable all Enterprise Features [EE]

---

• Release December 9th 2024
• Add a new option to the API endpoint POST /render/:idTemplate. The query parameter ?download=true can be used to download the rendered report directly.
  • When this option is used, the file is not stored on our servers. A single API call is enough to download the file. There is no need to call GET /render.
  • Important: The Content-Disposition header changes based on the HTTP status:
    • If the status is 200 (OK):

      Content-Disposition: 'attachment; filename="report.pdf"'
      Content-Type: 'application/pdf'

    • For all other statuses: The response remains in application/json format as before.

v4.24.2

• Release November 7th 2024
• [EE] Fix: Accept array filters with commas enclosed in quotes when using the :set() formatter with the prerelease tag {o.preReleaseFeatureIn=4022011} in ODT templates (enabled by default in Carbone v5).
• Fix: Prevent generating corrupted documents when a custom iterator contains a null or undefined value in a loop (e.g., {d[sort].id} -> {d[sort+1].id}). This fix is activated only if {o.preReleaseFeatureIn=4024000} (enabled by default in Carbone v5).

v4.24.1

• Release November 6th 2024
• [EE] Fix: the :html formatter in DOCX templates where the default template style might not have been applied.
• [EE] Fix: Accept array filters with commas enclosed in quotes when using the :set() formatter with the prerelease tag {o.preReleaseFeatureIn=4022011}.

v4.24.0

• Release October 25th 2024
• Accept xls (Excel 2003) template files only for document conversion.
• [EE] Fix the :html formatter in ODT templates where ordered and unordered lists contain mixed line break tags.
• [EE] Generate a fully compliant Factur-X PDF when the attached XML file is named 'factur-x.xml' (:attachFile('factur-x.xml')). It automatically detects the Conformance Level from the attached XML ('MINIMUM', 'BASIC WL', 'BASIC', 'EXTENDED', 'EN 16931') and adjusts the PDF metadata accordingly. Additionally, it sets SelectPdfVersion = 3 (PDF Version 3B) by default if not specified and if the output file is a PDF.
• [EE] Add a type parameter to the attachFile(filename, type) formatter to specify the PDF AFRelationship. Options for type include: "Source", "Data" (default), "Alternative", "Supplement", "Unspecified", "FormData", and "Schema".
• Close all pending TCP connections on exit to prevent indefinite waiting times.
• Fix: Allow :convCRLF to be used after :aggStrD and :aggStr.
• New formatter :preserveCharRef to preserve character references. By default, Carbone removes all forbidden characters before injecting data into XML (e.g., &, >, <, , etc.). As a result, injected character references like § ( = § ) are transformed into § in XML. This formatter prevents this transformation, preserving the original character reference. This feature is useful in specific XML generation scenarios where direct characters cannot be used (e.g., non-UTF-8 charsets) and the character reference must be retained. It accepts both numeric (e.g., d) and hexadecimal formats (e.g., 򡃯), in either lowercase or uppercase.

v4.23.7

• Release October 8th 2024
• [EE] Fixed a crash that occurred when the URL passed to appendFile or attachFile was not a string.

v4.23.6

• Release October 1st 2024
• [EE] Added support for bindColor in PPTX templates (for text, background color, and shapes). This feature is enabled when the flag {o.preReleaseFeatureIn=4022011} is included in the template.
• [EE] Fixed incorrect CSS attributes in charts created using Apache ECharts@v5a.
• Added en-be locale support for dates and numbers: DD/MM/YYYY, 100.000,80

v4.23.5

• Release September 19th 2024
• [EE] Fix escape characters &, >, and < for the aggregators :aggStr and :aggStrD.

v4.23.4

• Release September 17th 2024
• [EE] Support dynamic hyperlinks in headers and footers of DOCX templates
• [EE] Fix new :set() formatter when {o.preReleaseFeatureIn=4022011} is activated

v4.23.3

• Release cancelled

v4.23.2

• Release September 12th 2024
• New :t() formatter to translate text from JSON data. When the report is rendered, all text passed to the :t formatter is replaced with its corresponding translation found in a separate localization dictionary. Carbone automatically selects the localization dictionary that matches the lang attribute (en-us, en-gb, etc.). If a translation key is not found, Carbone prints the original text from the data in the final report (without translation).
• [EE] :attachFile formatter accepts URL with text/xml mimetype
• [EE] Fixes the :transform formatter in ODT/ODP when LibreOffice uses a different method to update the X/Y position of an item.
• [EE] :transform formatter support to move images in ODT/ODP/PPTX (only if {o.preReleaseFeatureIn=4022011} is activated)
• [EE] Optional flags {o.preReleaseFeatureIn=4022011} and {o.useHighPrecisionArithmetic=true} are removed from the generated document.
• [EE] Fix: The optional flag {o.preReleaseFeatureIn=4022011} is now detected before processing :set formatters.

v4.23.1

• Release September 2nd 2024
• Fixed :drop(shape) regression in PPTX templates when using native charts.

v4.23.0

• Release August 30th 2024

• New formatter :printJSON to print the content of an object/array. Example: {d:printJSON} prints the entire data object in the document.

• [EE] When using asynchronous jobs with carbone-webhook-url, optional headers carbone-webhook-header-X can be specified to customize the webhook headers sent. The X can be any key name. For example, if carbone-webhook-header-authorization: my-secret is included during report generation, the header authorization: my-secret will be added when Carbone calls the webhook.

• [EE] New formatter {d.pdf_or_xml_url:attachFile(filename)} to add one or multiple attachments to the generated PDF (e.g., Factur-X). This formatter accepts a single URL, can be positioned anywhere within a template, and does not produce any text output by itself.

  Carbone will return an error and will not generate the report in the following scenarios:

  • If the provided URL returns an error (with 2 retries for Carbone Cloud).
  • If the downloaded document is not a PDF or XML.
  • If more than 20 files are to be downloaded (On-Premise parameter: maxDownloadFileCount).
  • If the total size of all downloaded files exceeds 10 MB (On-Premise parameter: maxDownloadFileSizeTotal).
  • If the same document is included more than 5 times.

  Carbone will ignore the formatter and proceed to generate the report in the following cases:

  • If the provided URL is null, undefined, or an empty string.
  • If the final report is not in PDF format.

• [EE] Fix: The :color formatter did not apply the style correctly in HTML.

• ODS template supports :drop(sheet) and :keep(sheet) to show or hide a sheet

• XLSX template supports :drop(row) and :keep(row)

• Accept whitespace in attribute names using single quotes: {d.'new param'.'second level yes'[1].'sub obj'} if {o.preReleaseFeatureIn=4022011} is added in the template. Known limitation: not accepted inside formatters, array filters and aliases.

• Fix: Conditional blocks (show, hide) and smart conditional blocks (drop, keep) were not executed when combined with an array filter. In this example {d.arr[i, filter=3].id:ifEM:hideBegin} AAAA {d.arr[i, filter=3].id:hideEnd}, AAAA was printed if no rows matched the condition filter=3. This fix is available only when the tag {o.preReleaseFeatureIn=4022011} is activated.

• ⚡️ Transform your JSON without limits with :set(). This formatter, introduced in v4.5.2, was previously limited. Now it can be used to convert a flat JSON to a new hierarchical JSON with unlimited array depth, or the other way around. This feature is activated by adding the tag {o.preReleaseFeatureIn=4022011} and will be enabled by default in the upcoming Carbone v5.

  • :set(.attribute.id): Accepts a relative path to modify or add new attributes/objects inside an existing object.

  • :set(c.new[x=.y]): Creates or updates new arrays of objects.

    • c.new is the newly created array in c.. It can then be used with {c.new} as usual.
    • [x=.y] is an inner join expression. Only the = operator is accepted, and only exactly one expression is allowed between [] inside :set.
    • x is a relative path to the newly created object inside c.new. This expression cannot contain dots or square brackets.
    • .y is a relative path in the currently visited array (the tag before the :set expression). Use two or more dots to go up in the hierarchy if needed.

    How does it work? Before creating a new object inside c.new, Carbone checks if there is an existing object that matches the condition new.x = currentlyVisitedItem.y. If an object exists, it overwrites the existing object; otherwise, it inserts the new object inside the new array.

    This join expression works only on newly created arrays by Carbone, not on existing ones in the original JSON data passed when calling Carbone.

  • :set(c.new[x=.y].z): Creates or updates new arrays of objects and stores the result of the expression before :set in the child attribute z.

  • :set(c.new[x=.y].other[n=..m].o): Creates any nested array inside arrays with multiple join expressions.. Here Carbone creates two nested arrays new and other.

  • :set(c.new[x=.y].all[]): An empty join expression is accepted only for the last array in :set. In that case, all items are kept (including duplicated rows).

  Data:

    [
      { "country" : "A", "city" : "1A" },
      { "country" : "A", "city" : "2A" },
      { "country" : "A", "city" : "3A" },
      { "country" : "B", "city" : "1B" },
      { "country" : "B", "city" : "2B" }
    ]

  Result in {c.} with {d[].city:set(c.countries[id=.country].cities[].name)}:

    {
      "countries" : [
        {
          "id"     : "A",
          "cities" : [ 
            { "name" : "1A" },
            { "name" : "2A" },
            { "name" : "3A" }
          ]
        },
        {
          "id"     : "B" ,
          "cities" : [
            { "name" : "1B" },
            { "name" : "2B" }
          ]
        }
      ]
    }

  Result in {c.} with {d[]:set(c.countries[title=.country].cities[])}:

    {
      "countries" : [
        {
          "title"  : "A",
          "cities" : [ 
            { "country" : "A", "city" : "1A" },
            { "country" : "A", "city" : "2A" },
            { "country" : "A", "city" : "3A" }
          ]
        },
        {
          "title"  : "B" ,
          "cities" : [
            { "country" : "B", "city" : "1B" },
            { "country" : "B", "city" : "2B" }
          ]
        }
      ]
    }

  Result in {c.} with {d[].city:set(c.countries[title=.country].cities[])}:

    {
      "countries" : [
        {
          "title"  : "A",
          "cities" : [ "1A", "2A", "3A" ]
        },
        {
          "title"  : "B" ,
          "cities" : [ "1B", "2B" ]
        }
      ]
    }

v4.22.14

• Release August 23rd 2024
• Fix asynchronous rendering: allow multiple reports to be generated at the same millisecond

v4.22.13

• Release July 22nd 2024
• If config.factories = 0 (starting without the LibreOffice converter), the hardRefresh option is ignored.

v4.22.12

• Release July 22nd 2024
• [EE] Fix: :transform(axis, unit) for PPTX template did not work in certain cases
• [EE] Fixes studio hang when the template has no Carbone tags and adds an XHTML export option to get embedded images in HTML output

v4.22.11

• Release July 17th 2024

• [EE] The embedded stateless studio supports XML templates if the following conditions are met: Chromium-based browsers, output file type must be XML. The generated XML is indented and stored locally. It supports live reloads. If the template/data is updated, the generated XML is updated accordingly. VSCode, Sublime Text, or any editing tool can be used to update the XML template and see the result.

• [EE] Fixed :imageFit formatter for resizing image in ODS templates.

• [EE] Fix: Do not escape the character & two times in injected hyperlinks if {o.preReleaseFeatureIn=4022011} is active

• [EE] New :transform(axis, unit) formatter to move a shape on the X or Y axis:

  • axis can be x or y to move the item horizontally or vertically from its current position. Positive numbers = move to the right or bottom.
  • unit can be cm, mm, inch, or pt.

  The tag must be positioned inside the item to be transformed (alternative text or inside the shape itself). This tag prints nothing. Supported in ODP, PPTX and ODT templates. The feature is available only if the tag {o.preReleaseFeatureIn=4022011} is used. Be careful, if a loop is created between two shapes, the [i+1] tag must be in the shape that is higher in the layer order than the shape with the [i] tag.

v4.22.10

• Release July 4th 2024
• PPTX template support :drop and :keep formatters to control the visibility of elements:
  • Paragraphs p: insert the tag in a paragraph to show/hide: Text before {d.value:ifEM:drop(p)}. To delete multiple paragraphs simultaneously, specify the number of elements as a second argument: :drop(p, number).
  • Images img: use the tag in an image's alternative text or title to manage visibility: {d.url:ifEM:drop(img)}.
  • Shapes shape: place the tag in a text box or shape's alternative text to show or hide the shape: {d.showShape:ifEQ(false):drop(shape)}.
  • Tables table: insert the tag in a table cell to control the display of the table: {d.list:ifEM:drop(table)}.
  • Rows row: embed the tag in a table row to manage its visibility: {d.value:ifNE(20):drop(row)}. To delete multiple rows simultaneously, specify the number of elements as a second argument: :drop(row, number).
  • Charts chart: use the tag in the chart's alternative text to show or hide the chart: {d.list:ifEM:drop(chart)}.
• Fixed :html formatter: duplicating Bold or Italic tags is now supported

v4.22.9

• Release June 27th 2024
• Fix: When the tag {o.preReleaseFeatureIn=4022008} is used
  • ODS templates supports Carbone tags on the sheet name.
  • Accepts conditions on the i+1 part of a loop for backward compatibility.
  • Accepts XLSX with empty rows

v4.22.8

• Release June 21th 2024
• [EE] The formatter :color(part) accepts the new option part to color only a part of a paragraph (word, group of words, ...). The text or highlight color is applied within the same styled element. This feature is available only for ODT templates. Other template types will be supported on demand.
• New formatters:
  • :ceil(number) rounds towards closest higher integer number 3.5 -> 4 and -3.5 -> -3
  • :floor(number) rounds towards closest lower integer number 3.5 -> 3 and -3.5 -> -4
• Fix: A new engine now supports a wider range of use cases for conditional blocks and loops, thanks to an advanced algorithm from the upcoming Carbone v5. This example below, which combines filters, aggregators, and a conditional block on the same array, was not supported without this new algorithm: Also, it is not necessary to repeat the full tag at the end of a conditional block: {d:showEnd} or {d:hideEnd} is sufficient.

    {d.arr[sort>1].id:aggCount:ifGT(0):showBegin}
      {d.arr[i].id}
      {d.arr[i+1].id}
    {d:showEnd}

  ⚠️ This algorithm is enabled when the template contains the tag {o.preReleaseFeatureIn=4022008} or if the configuration / API contains preReleaseFeatureIn : 4022008. This tag means: "enable all pre-release features added up to v4.22.8".
• Fix: When the tag {o.preReleaseFeatureIn=4022008} is activated, Carbone improves the support of XLSX templates (no need to add empty cell)

v4.22.7

• Release June 14th 2024
• Update dependencies
• Improve Libreoffice detection on Arch Linux
• Fix: removed the possibility of prototype pollution in formatters
• Fix: showBegin/showEnd or hideBegin/hideEnd conditions in some rare cases when {o.preReleaseFeatureIn=4015000} is enabled.
  • When one condition ends exactly where a new condition starts and the XML does not contain other text
  • When conditions are used on text without XML, but there is some XML afterward
• Revert the :html evolution of v4.22.6 (v4.22.6 was not deployed in production)

v4.22.6

• Release June 7th 2024
• [EE] New feature for the :HTML formatter: the HTML is now supported in headers and footers of DOCX/ODT/PDF. Temporary exception for DOCX templates: if the HTML contains <img> tags, a image must be pre-inserted into the header/footer of the template, otherwise the image won't be displayed.
• [EE] Fix: more text editor styles applied to Carbone tags with the :html formatter are also applied to the injected HTML for DOCX/ODT templates. This fix addresses numerous spacing issues between paragraphs, before or after the injected HTML. List of new supported styles:
  • DOCX: Spacing between paragraphs, left indent, small caps, all caps, hidden text, double strikethrough.
  • ODT: Spacing between paragraphs, left indent, top/bottom margin, and text justify property

v4.22.5

• Release May 7th 2024
• Accepts filters on attributes of objects as if they were arrays:

  {
    myObject : {
      paul : '10',
      jack : '20',
      bob  : '30'
    }
  }

  In the report:

    {d.myObject[.att = jack].val} // 20
    {d.myObject[.val = 20].att}   // jack

  • use .att to access the attribute (paul, jack, bob)
  • use .val to print the value

v4.22.4

• Release April 18th 2024
• A new option {o.useHighPrecisionArithmetic=true} can be used within a template to activate arbitrary-precision decimal arithmetic for mathematical operations. When this tag is included in the template, the formatters :add, :sub, :mul, :div, :formatC, :formatN will maintain full decimal precision (Max 20). By default, and for all other formatters (such as aggregators, :abs, :mod, as well as mathematical expressions inside formatters and conditions), calculations uses double-precision floating-point numbers based on the 64-bit IEEE-754 standard.
• [EE] Fix bad CSS attributes in charts created with Apache ECharts@v5a

v4.22.3

• Release April 15th 2024
• Accept Microsoft Visio templates (.vsdx) with conversion to PDFs
• [EE] Fix drop(table) for ODP templates
• [EE] Add support of drop(item) for ODP and ODT templates to delete an item inside a list
• [EE] Fix crash when null values are passed in Apache Echarts options.
• [EE] Previously, charts created with Apache ECharts did not support setting custom font styles (size, family, weight). We have now resolved this issue with our own modified version of ECharts, optimized for SVG compatibility with our main document converter, LibreOffice. To ensure compatibility and prevent any impact on existing charts, please specify the ECharts version echarts@v5a instead of echarts@v5 when configuring your JSON. This @v5a version will become the default in Carbone v5.

  {
    "chartOptions" : {
      "type" : "echarts@v5a",
      "option" : {}
    }
  }

v4.22.2

• Release April 9th 2024
• Fix corrupted Excel file when aggSum:formatN is used since 4.22.0

v4.22.1

• Release April 8th 2024
• Fix: Aggregators may return incorrect results when used in combination with :set formatters (bug introduced by v4.22.0).

v4.22.0

• Release April 5th 2024

• The formatter convCRLF supports PPTX templates

• [EE] new formatter {d.pdf_url:appendFile} to append one or several PDFs at the end of the generated document. This formatter accepts either a single URL. It can be positioned anywhere within a template and does not produce any text output by itself. Carbone will return an error and will not generate the report in the following scenarios:

  • If the provided URL returns an error (with 2 retries for Carbone Cloud).
  • If the downloaded document is not recognized as a valid PDF.
  • If more than 20 files are to be downloaded (On-Premise parameter: maxDownloadFileCount).
  • If the total size of all downloaded files exceeds 10 MB (On-Premise parameter: maxDownloadFileSizeTotal).
  • If the same PDF document is included more than 5 times.

  Carbone will ignore the formatter and proceed to generate the report in the following cases:

  • If the provided URL is null, undefined, or an empty string.
  • If the final report is not in PDF format

• Improve file type detection when inserting images or PDF coming from external URL. It can read file extension in Content-Disposition if Carbone cannot find it Content-Type and URL.

• The execution order of multiple Carbone tags using the :set formatter is guaranteed within the same part of a document (body, header, footer, text box). As a result, you can create new variables that depend on previously created ones. When an existing set expression is used to create another set expression, the new expression must be define after or bellow the existing expression. Example:

    {d.price:add(10):set(d.total1)}
    {d.total1:add(20):set(d.total2)}
    {d.total2} = 30

• Supports variables with absolute JSON paths starting with d. or c. in mathematical formatters: add, sub, div, mul. For example: {d.total:add(d.val)}.

• Fix: Reduced the probability of generating corrupted XLSX files when using XLSX templates with :formatN to transform a JSON number into a native Excel number. Previously, if at least one injected value was a string, the generated XLSX file would become corrupted.

• [On-premise] Load custom formatters as Javascript in your Carbone instance:

  • Create a file named formatters.js under the plugin folder.
  • Write module.exports = { }, then insert your Javascript function between curly brackets: One function is equal to one formatter.
  • Here is the minimum code of a formatter:

    function addText (d, text) {
      return d + text;
    }
    /**
     * Details:
     * "d" the tag value
     * "text" the first argument of the formatter
     * A value must be returned otherwise it will print nothing in the document.
     * Example usage: {d.value:addText(' euro')}
     */ 

    Find formatters examples on the following page: https://github.com/carboneio/carbone/blob/master/formatters/string.js

• Accept to send a volatile template when calling the API POST /render/template. This template is never stored and it does not trigger the middleware readTemplate

    {
      data      : {},
      template  : "base64-encoded-file",
      convertTo : "pdf"
    }

• If options.isDebugActive = true. POST /render returns internal metrics information in debug sub-object. All Time values are in microsecond.

  {
    debug    : {
      metrics : {
        preProcessTime  : 0,
        planTime        : 0,
        mergeTime       : 0,
        concatTime      : 0,
        fetchImageTime  : 0,
        fetchImageBytes : 0,
        fetchFileTime   : 0,
        fetchFileBytes  : 0,
        postProcessTime : 0,
        convertTime     : 0,
        renderTime      : 0,
        batchSize       : 1
      }
    }
  }

v4.21.0

• Release Mars 16th 2024
• [On-premise] Studio Light: Added button to copy the Template ID of the previewed template.
• [EE] Fixes support for SVG image replacement in DOCX templates, especially when using LibreOffice 24 for document conversion.
• Fixes broken DOCX template with bi-directional loops when the template is created with Google Docs
• Fixes the row height in spreadsheets. It forces LibreOffice to calculate the optimal row height if the row style in ODS templates has the style:use-optimal-row-height='true' parameter.

v4.20.0

• Release February 26th 2024
• [EE] Support batch processing of multiple documents (max 5000) with one API call. The result is a zip file of documents. This feature is temporarily limited to some customers. Please contact us if you need to use it. Rate-limit : one batch processing at a time by tenant.

v4.19.0

• Release February 22th 2024
• [EE] The :color formatter supports updating the background color of a page with :color(page, background) options in DOCX templates.
• [EE] New aggregator formatters:
  • :aggCountD : Count the number of distinct values. Null or undefined values are ignored.
  • :aggStrD : Aggregate distinct values. Null or undefined values are ignored.
• The :formatC(precision, currency)' formatter accepts an optional second parameter to force the target currency. It overrides the globalcurrencyTarget` option.
• If the global option currencySource is undefined, no conversion is done when using the formatters convCurr and formatC.
• Fixes locales fr-CH and rm-CH. Thousand separator is ' and decimal separator is . like de-CH and it-CH.
• The maximum number of repetitions when using this syntax {d[i+1*qty] can be set globally with the maxRepetitionFactor parameter (400 by default).
• [EE] Update to latest Node v18

v4.18.0

• Release February 14th 2024
• [EE] Support dynamic pictures into Shapes for DOCX template only. Create images with rounded corners, or images with special shapes.
• Fixes formatter substr(begin, end, wordMode) when word mode is active. If wordMode=true, it never cuts words. It can be used to print successive lines of text of constant width. The previous algorithm of v4.12.0 could create some gaps (missing words) when doing successive calls of substr. Here is the problem:
  • Each call to substr in the template is calculated independently.
  • But the line of text we want to print depends on all the lines of text that have been printed before. To solve this, each subsequent call to substr must obey these constraints:
    • A constant line width (end - begin) must be used between successive calls of substr to print all the words of the text without gaps. For example:
  • {d.text(0 , 50 , true)} -> line 1 of 50 characters
  • {d.text(50 , 100, true)} -> line 2 of 50 characters
  • {d.text(100, 150, true)} -> line 3 of 50 characters
  • {d.text(150, 200, last)} -> line 4 of infinite characters
    • last can be used instead of true to print the rest of the text, even if it is longer than the defined line width.
    • A word can only be truncated if it does not fit in the line. In this case, the word always starts at the beginning of a new line.
• ⚡️ Carbone On-Premise starts with Community Edition features if no valid license is provided. It collects some statistics at startup (version, subscription). These statistics can be deactivated.
• Improve the stability of reports based on ODT templates that include native charts.

v4.17.0

• Release January 30th 2024
• To include a single-quote character in formatter parameter, write two adjacent single quotes, e.g., 'David''s Car'. Note that this is not the same as a double-quote character (").
• carbone.renderXML accepts html templates if option.extension is html
• Support drop in ODS (img, row) and HTML (table, p, row) templates.
• ⚡️ Support keep, the opposite of drop
• [EE] 🤩 New features and new supported templates for the new :color(scope, type) formatter
  • Supported in ODT, ODP, HTML and DOCX templates
  • scope can be p (by default), cell, row, shape to apply color to the current paragraph, cell, row or shape
  • type can be text (by default), highlight for text, background for cells, rows and shapes, border for shapes only
  • accepts only 6-digit hex color notation with or without hashtag, lowercase or uppercase: #FF0000 or FF0000. Carbone replaces wrong color values with light gray (#888888)
  • for HTML, the color is applied with the style attribute only to the selected HTML tag (tr, td and p)
  • This new method is better than the old method with bindColor:
    • It is much easier to use. The color is only applied to the targeted scope where the carbone tag is (like :drop or :keep).
    • It does not break other styles of the targeted element. New colors are merged with existing style (e.g. cell border, ...).
    • Supports different conditions on text and highlight at the same time for ODT/ODP. For example: {d.col:color(p)} {d.nb:ifEQ(true):show(FF0000):color(p, highlight)} is possible.
    • Manages priority between overlapping areas. A color applied to the current paragraph has a higher priority than a color applied to the current shape, cell and row.
  • WARNING: There are known limitations in this version:
    • in ODP (text, table, and shapes) and ODT (shapes only), the :color formatter requires that a non-default style is already applied to the target element in the template (for example, a custom text color).
    • complex nested tables with colors on sub-tables are not fully supported.
    • highlight is not managed in docx template
    • Cannot be used with aliases
  • Fixed :color(scope, type) bug for DOCX templates when the Carbone tag :color tag is quite long with many conditions for example (v4.16.0)
• 🌈 New :aggStr(separator) formatter, a more powerful version of :arrayMap. By default the separator is ', '.
  • d.user.phones[]:aggStr() -> +331112345678, +331112345678, +331112345678
  • d.cars[electric=true].brand:aggStr(' - ') -> Tesla - BYD - Kia

v4.16.2

• Release January 3rd 2024
• Increase glyphs cache size of Libreoffice at startup to speed up the PDF conversion of documents with 50 pages of table/text (potential gain: x10)

v4.16.1

• Release December 28th 2023
• Fix array of string access in a loop of another array
• The formatter :set(c.value) accepts to store values in complement object c.

v4.16.0

• Release December 19th 2023
• Fix: accept undefined formatOptions when convertTo is an object
• [EE] New :color(scope, type) formatter only for DOCX template: {d.myHexColor:color(p)}
  • scope can be p (by default), cell, row for applying color on the current paragraph, cell and row.
  • type can be text (by default), background for cells and rows.
  • accepts only 6-digit hex color notation with or without hashtag, lower case or upper case: #FF0000 or FF0000. Carbone replaces incorrect color values with white (#FFFFFF)

v4.15.7

• Release December 14th 2023
• New aggregator :cumCountD to count the number of distinct values

v4.15.6

• Release November 16th 2023
• [EE] Fix :html formatter did not remove forbidden characters such as  in DOCX, ..., XLSX, ODT
• Fix weird behavior when a tag with direct array access is placed on the same row as the loop:

     {d.list[0].name} {d.list[i].name}
                      {d.list[i+1].name}

• [EE] On-Premise: new feature to remove macro of ODS/ODT templates
• [EE] Update echarts to 5.4.3

v4.15.5

• Release October 18th 2023
• [EE] Before v4.15.5, Carbone executed the job which deletes unused reports (= not downloaded) every 60 minutes. The default retention time was also 60 minutes. Carbone could keep a generated report during 1 hour and 59 minutes 59 seconds instead of 1 hour (time announced in our documentation). Now, Carbone executes the cleaning process every 10 minutes, and the retention time is 50 min. As a result, a generated report is available for download for 50 minutes to a maximum of one hour.
• [EE] On-premise: accept both Basic authentication and JWT when basic authentication is enabled
• [EE] On-premise: add the possibility to set the scheduled job time interval for cleaning unused templates and unused generated reports. Set the environment variable CARBONE_EE_CLEANINTERVALTIME, or --cleanIntervalTime CLI options, or set cleanIntervalTime on the configuration file. The value must be in minutes (minimum : 1, maximum : 35791). The default value is "every 10 minutes".

v4.15.4

• Release October 13th 2023
• [EE] Fix the :html formatter: the default font size from a DOCX template is always applied on the rendered HTML.
• [EE] Accept charts in PPTX without style/color description
• [EE] Fix a minor PPTX generation issue when using chart

v4.15.3

• Release October 6th 2023
• [EE] On-Premise: GET / returns the same response as GET /status when the embedded studio is disabled

v4.15.2

• Release October 3rd 2023
• [EE] Fix Basic authentication regression for On-premise version

v4.15.1

• Release September 26th 2023
• Fix: Accepts to access adjacent array with ... Example: {d.array[i]..otherArray[condition=other].id}
• [EE] Improve embedded v5 engine (see v4.15.0) when the formatter drop is used on i+1 part of an array
• [EE] Dynamic colors are supported into ODP documents. Only the text and background colors can be changed dynamically. Changing shapes and cell background are not supported.

v4.15.0

• Release September 20th 2023
• [EE] Includes a part of the new highly reliable engine of Carbone v5. It supports more use cases and reduces virtually to 0% the probability to have "Could not open document" error. The new engine accepts syntaxes that are not logical for a computer but acceptable for a human, especially in a document. For example, writing tags in a document can generates such an algorithm sometimes:

    IF
      START LOOP
        END IF
      END LOOP
    # END IF should be here for a computer. Carbone fixes it automatically

  This v5 engine is only enabled if the template contains the tag {o.preReleaseFeatureIn=4015000} or if the configuration / API contains preReleaseFeatureIn : 4015000. This tag means: "enable all pre-release features added up to v4.15.0".

v4.14.4

• Release September 12th 2023
• [EE] The :imageFit formatter is now supported for PPTX documents.

v4.14.3

• Release August 28th 2023
• [EE] Fix dynamic images mixed with dynamic hyperlinks for DOCX documents
• [EE] On-premise: change the message "Unknown parameter" to "Additional plugin parameters detected" when Carbone starts.
• [EE] Fix corrupted PPTX files when the template uses multiple charts on many slides
• [EE] When converting ODT/DOCX to HTML, embed images in the HTML instead of creating external links (works only with xhtml export)

v4.14.2

• Release August 4th 2023
• [EE] Fix corrupted documents when ODP/ODT templates contain charts in some rare cases.

v4.14.1

• Release August 1st 2023
• [EE] Fix charts bug in Docx templates since 4.14.0

v4.14.0

• Release July 26th 2023
• [EE] Support two new template formats:
  • XML Word 2003 can be used to generate: pdf, txt, docx, odt, doc, jpeg, png
  • XML Excel 2003 can be used to generate: pdf, csv, ods, xlsx, xls, jpeg, png
  • Known limitations: enterprise features are not supported for the moment (images, barcode, colors, html, etc...)
• [EE] Fix POST /render/:templateId: when a body request is larger than the maxDataSize (Carbone CLoud API limit is 60MB), a 413 code is returned instead of 500, and the error message is Content too large, the JSON size limit is X MB instead of PayloadTooLargeError.
• [EE] Fix :html formatter for ODT/DOCX/PDF documents:
  • Empty paragraphs on the generated document are now keeping the style applied on the template.
  • HTML entities: Support unicode values for all currency symbols and most used punctuation. For instance ' will print '.
  • Support the rendering of nested ordered/unordered lists with mixed paragraphs and anchors tags.
• [EE] Support charts in ODP and PPTX templates (same chart types as DOCX/ODT)
• Fix wrong template format detection when a PPTX contains a chart which an embbedded XLSX file

v4.13.0

• Release July 7th 2023
• [EE] Carbone API status codes fixes:
  • For all endpoints, if the API key is missing or not correct, a 401 code is returned instead of 500.
  • POST /template: if the template attribute is missing, a 422 code is returned instead of 400.
  • POST /template: if the template format is not supported, a 415 code is returned instead of 200.
  • POST /render/:templateId: if the data attribute is missing, a 422 code is returned instead of 400.
  • GET /render/:renderId: if the generated document does not exist, a 404 code is returned instead of 200.
  • DELETE /render/:templateId: if the template does not exist, a 404 code is returned instead of 400.
• Fix condition ifNIN when the data is null or undefined
• Fix corrupted document when html formatter is used and the text contains &D;

v4.12.2

• Release July 3rd 2023
• Fix chained formatters :split():arrayJoin('', 2, 3)
• Fix formatter substr
• Array filter: accept to filter null or undefined values

v4.12.1

• Release June 28th 2023
• Fix ellispis -> ellipsis formatter

v4.12.0

• Release June 28th 2023
• [EE] Carbone On-Premise: add parameter maxTemplateSize (bytes) to change default limit (20MB)
• [EE] Fix GET /template: If the template doesn't exist, the statusCode 404 is returned instead of 400.
• [EE] Fix POST /template: If the template is too large, the statusCode 413 is returned instead of 400
• Add new fomatters:
  • ellipsis(maxLength): add three dots ... if the text is longer than maxLength
  • substr(begin, end, wordMode): add a third option to cut the text without cutting a word if wordMode = true (false by default)
  • abs(): get the absolute value of a number
  • replace(oldText, newText): replace a text by another
  • split(delimiter): split a text using a delimiter and generate an array
  • prepend(text): add a prefix
  • append(text): add a suffix
  • arrayJoin(separator, index, count): add two options to cut the array and select only a part of it

v4.11.2

• Release June 14th 2023
• [EE] On-Premise: New option to enable more security controls, following ANSSI (France) and BSI (Germany) recommendations. Contact us for more information.
• [EE] All APIs returns 400 Bad request instead of 404 Not Found when idTemplate is not valid

v4.11.1

• Release June 2nd 2023
• Accept image from dropbox with mime type application/binary and image extension in capital letters

v4.11.0

• Release June 1st 2023
• Support drop formatter in headers and footers of DOCX and ODT documents.
• :drop(h) can be used to delete heading elements for ODT templates only. On LibreOffice, it relates to heading style 1, 2, 3, 4 and custom styles. For DOCX templates, :drop(p) must be used to delete heading elements.
• Fix: using :html or :drop with an empty value inside a table cell no longer creates a corrupted document with DOCX templates.
• Fix: sometimes drop formatter was not found and not executed by Carbone
• Fix HTML formatter: The style of <ul> or <ol> lists are now correctly rendered into DOCX documents.
• Fix a random dynamic chart issue with some DOCX templates
• Fix chart distortion issue when converting DOCX to PDF if the document contains more than 20 charts
• Fix hyperlink error when the value is not a string

v4.10.6

• Release May 12th 2023
• [EE] Fix: accepts whitespace in image URL like before v4.10.2 even if it is not recommended

v4.10.5

• Release May 8th 2023
• [EE] set default webhook timeout to 16 seconds (it was 3s since 4.10.0)

v4.10.4

• Release May 4th 2023
• [EE] Egress traffic: When a request retry is attempted, do not use the egress proxy again if the primary egress proxy is configured and the secondary egress proxy is undefined

v4.10.3

• Release May 3rd 2023
• [EE] Improve automatic HTTP request retry of images with the last version of rock-req

v4.10.2

• Release May 3rd 2023
• [EE] Fix options.isDebugActive = true: do not return internal (not visible by user) tags
• [EE] Fix crash when dynamic image URL contains forbidden characters
• [EE] Fix crash when the socket is broken while the stream is still in progress for GET /template/:templateId and GET /render/:reportId
• [EE] Fix HEAD /render/:renderId: The request don't delete the generated document anymore.
• [EE] Fix GET /template/:templateId: If the template doesn't exist, the statusCode 404 is returned instead of 400.

v4.10.1

• Release April 17th 2023
• [EE] Disable egress proxy by default

v4.10.0

• Release April 17th 2023
• [EE] Support vector barcodes with the new svg option to improve print quality: {d.number:barcode(qrcode, svg:true)}

v4.9.2

• Release April 14th 2023
• [EE] Fix dynamic hyperlinks when d is an array for DOCX templates

v4.9.1

• Release April 12th 2023
• [EE] with XLSX/ODS templates: formatN converts values to native Excel number even if d is an array. formatN without parenthesis is also accepted.
• [EE] On-Premise: Add new options to use a proxy for egress traffic
• Fix dynamic hyperlinks error URL and add slash to URLs

v4.9.0

• Release March 3rd 2023

• Accept absolute path which starts by d. or c. without quotes in formatters like this : {d.id:print(d.other)} or {d.id:print(c.other[0].test)}

• [EE] Force download of the rendered report when the query parameter ?download=true is set

• [EE] Increase request timeout from 5 to 6 seconds to download images from URLs

• Accept loops on array of arrays (unlimited number of nested array) like this:

    {d.myArray[i][i].val}
    {d.myArray[i][i+1].val}
    {d.myArray[i+1].val}

• Support aggregators on array of numbers. For example:{d.table[]:aggSum} = 63 if table contains [10, 20, 33]. Known limitation: Array filters on array of strings/numbers are still not supported.

• ⚡️ Support loops on array of string/numbers. Example:

  Data:

    {
      "myArray" : ["yellow", 125, "blue"],
    }

  Template:

    Direct access: {d.myArray[0]:upperCase()}
    Loop:
      - {d.myArray[i]}
      - {d.myArray[i+1]}

  Result:

    Direct access: YELLOW
    Loop:
      - yellow
      - 125
      - blue

  To maintain backward compatibility with existing templates, if the template contains at least one attribute access on myArray (Example: {d.myArray[i].subObjectAttribue}), Carbone considers myArray is an array of objects and it disables this feature. In this case, {d.myArray[i]} prints nothing and is neutral for array filters even if the array contains a printable type like a string or a number, as it was before v4.9.0.

  Data:

    {
      "otherArray" : [
        {"id" : 10},
        {"id" : 20},
        125,
        "blue"
      ]
    }

  Template:

    Without filters:
      - {d.myArray[i]} {d.myArray[i].id}
      - {d.myArray[i+1]}
    With filters:
      - {d.myArray[i]} {d.myArray[i, id>9].id}
      - {d.myArray[i+1]}

  Result:

    Without filters:
      -  10
      -  20
      -
      -
    With filters:
      -  10
      -  20

  Known limitation: Array filters on array of strings/numbers are still not supported.

• [EE] Includes a part of the new v5 engine that supports more use cases in some complex templates. This v5 engine is only enabled if the template contains the tag {o.preReleaseFeatureIn=4009000}. This tag means: "enable all pre-release features added up to v4.9.0". Please only use this tag if we tell you to use it on our live chat.

v4.8.3

• Release February 15th 2023
• [EE] Experimental: increase the limit to 400 repetitions maximum when using the repetition feature {d[i+1*qty]

v4.8.2

• Release February 7rd 2023
• Fix dynamic image injection in some XLSX files

v4.8.1

• Release February 3rd 2023
• Fix regular hyperlinks in DOCX (bug introduced by v4.6.3)

v4.8.0

• Release February 1st 2023
• [EE] BREAKING CHANGE in Carbone On-Premise plugins: the callback of getPublicKey must take two arguments (err, publicKeys). If the first argument contains an error, the bad token is kept in quarantine area for 60s instead of infinitely.
• [EE] Update Echarts to 5.4.1, dayjs to 1.11.7
• Fix horizontal loops in DOCX tables
• Ignore hardRefresh : true if the output file type is the same as the template type, and the file type is unknown by LibreOffice (example: XML to XML)
• Fix: The parameter of a formatter should not be considered as a dynamic variable if it starts with a dot and is surrounded by quotes as in v3.x

v4.7.0

• Release January 26th 2023
• Improve error message when there is a missing [show|hide]Begin/End
• Fix: Accept PDF password with integers
• [EE] Support Bubble charts in Docx templates

v4.6.8

• Release December 7th 2022
• Fix array filters when the same filter is used with different operators in multiple tags. Ex {d[type=ok].id} {d[type!=ok].id}

v4.6.7

• Release December 5th 2022
• Fix regression of v4.6.6 when multiple loops are used on the same table.

v4.6.6

• Release November 30th 2022
• Fix bugs when using two Carbone tags which are using similar paths in some complex situation (d > obj > arr > obj > [arr|obj] > att). Very hard to explain here!
• Increase max downloaded image in parallel from 5 to 15.

v4.6.5

• Release November 25th 2022
• Accept parentheses in formatters between single quotes

v4.6.4

• Release November 17th 2022
• Fix aspect ratio of images if the same image is used twice in a second section of a document

v4.6.3

• Release November 10th 2022
• Full support of table of content and bookmarks (internal hyperlinks) in DOCX templates

v4.6.2

• Release November 4th 2022
• Fix: Encode HTML special characters when generating SVG with Echarts

v4.6.1

• Release October 28th 2022
• Fixed :drop(row, nbrToDrop) for DOCX documents: bookmarked table rows can be deleted.
• negative numbers can by used when filtering with the iterator i
  • {d.arr[i, i<-2].id} {d.arr[i+1, i<-2].id} print all elements except the last 2 items
  • {d.arr[i, i<-1].id} {d.arr[i+1, i<-1].id} print all elements except the last item

v4.6.0

• Release October 18th 2022
• Added the option table for the :drop(element) formatter to delete table conditionally. Available for DOCX, ODT and ODP templates. Usage: {d.value:ifEM:drop(table)}. The tag must be located within a table cell.
• ⚡️ Added for PPTX templates:
  • Dynamic images are supported. Set the Carbone tag into the image alternative text. It is not possible to create a loop of images.
  • All barcodes are supported. Set the Carbone tag into the image alternative text and chain the :barcode(type), such as creating a QR Code: {d.productCode:barcode(qrcode)}.
• Add user-agent Carbone when webhook is called. Print HTTP status webhook response in logs.
• Added types for the ifTE(type) formatter: boolean, binary, array, object, number, integer. Usage: {d.value:ifTE(number):show('It is a number!')}, the conditional formatter checks if the type of the value is a number.

v4.5.2

• Release October 9th 2022
• ⭐️ Add the possibility to store values in data with new formatter :set(d.value). Known limits:
  • accepts only to store values at the root level of d, and only in d.
  • cannot be used in aliases
• Fix aggregators aggSum, aggAvg, aggMin, aggMax, aggCount when used with filters (without iterators) like {d.cars[id=1].wheels[size=100].makers[].qty:aggSum}
• New Formatters mod() to compute modulo
  • Example: {d.val:mod(2)} returns 0 if d.val = 4

v4.5.1

• Release October 7th 2022
• Added for the :drop(element) formatter:
  • ODP templates are supported
  • A new argument slide is available for ODP template only to delete slides. Usage: {d.value:ifEM:drop(slide)}.
  • drop(p, nbrParagraphs) accepts a second parameter nbrParagraphs to delete multiple paragraphs at once. For instance {d.data:ifEM:drop(p, 3)}, meaning the current and next two paragraph will be removed if the condition is validated. By default, the formatter :drop(p) hides only the current paragraph.
• aggregators (aggSum, aggAvg, ...) convert string to floats. null or undefined values are converted to
  • aggSum : 0
  • aggMin : +infinity
  • aggMax : -infinity
  • aggAvg : 0

v4.5.0

• Release October 5th 2022
• Dynamic parameters passed in formatters with a dot . accepts dynamic array access between brackets [.i]: It can be used to select the corresponding element of another array: {d.myArray[i]:myFormatter(..otherParentArray[.i].id)}. [.i] matches with the index of myArray[i]. Multiple dots ([..i], [...i]) can be used to access other parents arrays. If the index goes above otherParentArray length, it will return an empty string
• Fix: HTML comment tags are not rendered and skipped by :html formatter
• Fix: barcodes accepts integer values
• On-Premise:
  • Improve logs messages for webhooks
• Performance: huge gain from x10 to x20 when Carbone builds the final result of the report before the conversion

v4.4.1

• Release September 26st 2022
• On-Premise:
  • add options xlsmEnabled to accept export to xlsm format (false by default)

v4.4.0

• Release September 21st 2022
• Add new formatter ifTE(string) to test if a value is a string. Only "string" is supported for now.
• Fix: In docx templates, array filters could be ignored when the loop includes images (bug introduced in v4.1.0 to fix broken docx)
• On-Premise version:
  • Improve error message of HTTP API /template /render
  • Support execution of on Windows (beta)
• New: Compute the difference between two dates with formatter: d.fromDate:diffD(toDate, unit, patternFrom, patternTo).
  • fromDate and toDate can be ISO 8601 format or any format defined with patternFrom and patternTo
  • unit can be millisecond(s) or ms, second(s) or s, minute(s) or m, hour(s) or h,year(s) or y, month(s) or M, week(s) or w, day(s) or d, quarter(s) or Q. Here are examples with d.fromDate = 20101001:
  • {d.fromDate:diffD(20101201, days)} => 61
  • {d.fromDate:diffD(20101201, hours)} => 1465
• Fix: formatD ignores the timezone if only a date is parsed without the time (without hour, minute, second). Example: when the timezone is america/guayaquil in options.timezone
  • '2010-12-01':formatD(LL) returns December 1, 2010. Before this version Carbone returned November 30, 2010
  • '2010-12-01T01:00:00Z':formatD(LL) returns November 30, 2010 8:00 PM

v4.3.0

• Release September 12st 2022
• hide formatter becomes drop to avoid confusion with hideBegin/hideEnd/show. hide was introduced in 4.2 and is still not offically documented.
• drop(row, nbrRows) accepts a second parameter to select the number of row to remove
• drop formatter can't be used into Carbone aliases.

v4.2.0

• Release September 8st 2022
• Fixed parsing of Carbone tags when empty string are used between two single quotes. Ex. {d.text:print(''):print('HIGK LMN')} prints HIGK LMN instead of HIGKLMN
• Improved error messages if a square bracket is used in array accessor [...]
• Added: Accepts dynamic parameters in array filters, with infinite path depth. Example: Data

      {
        parent : {
          qty : 5,
          arr : [{
            id : 1
          }]
        },
        subArray : [{
          text : 1000,
          sub : {
            b : {
              c : 1
            }
          }
        }]
      }

  • {d.subArray[.sub.b.c = 1].text}: filter using infinite object path depth. Alternative syntax without a dot is also accepted ({d.subArray[sub.b.c = 1]}) for backward compatibility
  • {d.subArray[i = .sub.b.c].text}: filter using dynamic values on the right operand
  • {d.subArray[i = ..parent.qty].text}: filter using dynamic values from parent objects on the right operand
  • {d.subArray[.sub.b.c = ..parent.qty].text}: filter using complex path for the left and right operand in the same time This feature is really powerful but some syntax are not supported yet:
  • {d.subArray[i = .i].text}: using .i to join two arrays is not supported
  • {d.subArray[i = ..parent.arr[0].id].text}: accessing a specific array element is not supported
  • {d.subArray[i = ..parent.arr[.i].id]text}: accessing a specific array element according to the current iterator is not supported
• [EE] Added hide conditional formatter for DOCX/ODT/PDF to hide document elements: images, paragraphs, table rows, shapes and charts. The rendering is always accurate and simpler to use compared to hideBegin/hideEnd or showBegin/showEnd. The first argument passed to :hide(argument1) is the element to hide, it can be:
  • p to hide paragraphs, usage: {d.text:ifEM:hide(p)}. The tag must be inside a paragraph. Every elements inside the paragraph are also removed if the condition is validated.
  • row to hide a table row, usage: {d.data:ifEM:hide(row)}. The tag must be inside a table row. Every element inside the row are also removed if the condition is validated.
    • Option nbrRowsToHide: Set the number of rows to hide as a second argument {d.data:ifEM:hide(row, nbrRowsToHide)}, such as: {d.data:ifEM:hide(row, 3)}, meaning the current and next two rows will be removed if the condition is validated. By default, the formatter :hide(row) hides only the current row.
  • img to hide pictures, usage: {d.img:ifEM:hide(img)}. The tag must be included within the image' title, description or alternative text.
  • chart to hide charts, usage: {d.dataset:ifEM:hide(chart)}. The tag must be included within the chart' alternative text.
  • shape to hide shape (square, circle, arrows, etc...), usage: {d.dataset:ifEM:hide(shape)}. The tag must be included within the shape' title, description or alternative text.
• [EE] Improved: ODS templates support loops of dynamic images. Setting the image anchor "To cell" is required.

v4.1.0

• Release August 22st 2022
• [EE] New: convCRLF prints \n and \r\n as new lines in ODS template instead of strings
• [EE] New: On-Premise Embedded Studio features:
  • Drag and drop a JSON file and the studio automatically updates current left panel (data, complement, enum or translation)
  • Drag and drop a template file and the studio automatically uploads the template and updates the preview
  • Add HTML export
• New: interval/duration formatters: d.duration:formatI(patternOut, patternIn). It accepts duration in milliseconds (by default), or ISO format (ex. P1Y2M3DT4H5M6S).
  • patternOut and patternIn can be millisecond(s) or ms, second(s) or s, minute(s) or m, hour(s) or h,year(s) or y, month(s) or M, week(s) or w, day(s) or d.
  • patternOut can be human and human+. Here are examples with d.duration = 3600000:
  • {d.duration:formatI(ms)} => 3600000
  • {d.duration:formatI(s)} => 3600
  • {d.duration:formatI(minute)} => 60
  • {d.duration:formatI(hour)} => 1
  • {d.duration:formatI(human)} => an hour
  • {d.duration:formatI(human+)} => in an hour
  • {d.duration:formatI(hour, second)} => 1000
• [EE] :html formatter updates:
  • New: the image tag <img> is supported and rendered into DOCX/ODT/PDF documents.
    • The image source attribute can be an URL or Data-URL, such as <img src=""/>
    • The image size is rendered based on width and height attributes provided by the HTML tag, such as <img src="" width="300" height="100"/>. Values must be pixels. If width or height attributes are missing, the size of 5cm (1.96in) is applied by default while retaining the image aspect ratio.
  • New: The HTML content can now be rendered into "heading" styled text on your text editor.
  • Fixed: Paragraph spacing are now rendering correctly (e.g. <p> <ul> <li>content </li> </ul>, <p><p> <p>content)
  • Fixed for ODT templates: Hyperlinks tags inside lists are now rendered without errors.
  • Fixed for DOCX templates: ordered and unordered lists size the same as the text
• [EE] Fixed dynamic hyperlinks with query parameters for ODT templates
• [EE] Fixed broken Docx files when shapes were duplicated by Carbone

v4.0.0

• Release June 25st 2022
• [EE] ⚡️ Main features summary (see v4.0.0-alpha.0 for details)
  • Support dynamic charts in LibreOffice and Word + echarts
  • Improved debug message output
  • New aggregator formatters: aggSum, aggAvg, aggMin, aggMax, aggCount
  • Accept more complex parameters in formatters: expression with arrays, Mathematical formulas (v3.5.0)
  • Improved On-Premise Embedded Studio: multi-language, change export file format, automatic JSON generation from template
  • Accept formatters after conditional formatters (v4.0.0-beta.1). For example, bindColor can be used with conditions
• [EE] Fix chart in DOCX when there is no loops (filtered array)
• [EE] Fix stateless studio crash when template does not contain any Carbone tags
• [EE] 🌈 bindColor formatter replaces background and line colors of shapes in DOCX only.
  • The bindColor tag must be written in the document (NOT in alt text of the shape)
  • The replaced color in the template must be RGB. Select "RGB sliders" tool to defined the color in MS Word.
• [EE] Use lossless image compression by default to speed up PDF rendering and improve image quality
• [EE] Remove experimental support of images in HTML with :html formatter for ODT template added in v4.0.0-beta.3 (postpone in 4.1)
• [EE] Barcode improvements (dependency update):
  • The horizontal alignment of text in matrix symbols was fixed.
  • Various fixes were made for the encoding of Data Matrix, DotCode and Micro QR Code symbols.
  • The encoding of QR Code symbols was optimized.
  • The encoding of Rectangular Micro QR Code symbols was aligned with the final release of the specification.
  • The linear render now uses filled polygons rather than stroked lines.
  • Code 93 Extended was amended to not shift encode "$%+/" symbols.
  • Support was added for USPS FIM E marks.
  • Support for AI (715) was added to the GS1 linter.
  • Ultracode tile colours are now defined as RGB rather than CMYK. New tile colour patterns are defined for the upcoming revision.
  • A bug in the encoding of certain Aztec Code symbols was fixed.
  • A bug in the encoding of certain Dotcode symbols was fixed.
  • A bug in the encoding of QR Code symbols containing Kanji compression was fixed.
  • The rMQR encoding was optimised, potentially resulting in smaller symbols.
  • The colours for Ultracode symbols were changed to RGB values rather than CMYK.
  • The metrics for Ultracode symbols was updated and a raw mode was added.

v4.0.0-beta.2

• Release June 1st 2022
• Fix crash with very complex JSON map

v4.0.0-beta.1

• Release May 25th 2022
• [EE] Accept formatters after conditional formatters. It solves many issues, such as dynamic colors with conditions:
  • {d.value:ifLT(10):show(0):formatN} : formatN works even if the condition ifLT(10) is true
  • {bindColor(fde9a9, hexa) = d.value:ifLT(10):show(FF00FF):ifLT(20):show(005FCF):elseShow(FFDD00)}: conditional colors works!
• Fix multiple reDoS and optimize parsing of some templates
• [EE] Include 3.5.2
• [EE] Dynamic chart:
  • Fix crash when DOCX/ODT templates contain empty files
  • Fix bad behavior when ODT template contains images with dynamic charts
  • Fix chart binding when values contain white spaces
  • Fix ODT charts when images are used for background

v4.0.0-alpha.1

• [EE] BREAKING CHANGE: the specific tag {bind becomes {bindChart. Example: {bindChart(91) = d[i].valCol1}
• [EE] DOCX Charts improvements
  • Manage loops to repeat multiple charts in DOCX template made by MS Office
  • Update embedded spreadsheet
  • Supports only Column, Line, Pie charts
  • Carbone tags must be written with all i and i+1 rows and columns in related Excel spreadsheet.
  • Using the specific tag {bindChart is not mandatory for DOCX because MS Word accepts Carbone tags in chart values
• Fix crash when a condition is used just before a filtered loop

v4.0.0-alpha.0

WARNING: Native charts in LibreOffice and Word still need a lot of work before being stable for production

• [EE] ⚡️ Carbone supports dynamic charts with two methods:

  • 1 - using native charts of MS Word or LibreOffice
  • 2 - or using Apache ECharts 5.3.3 object descriptors to generate advanced cha2

  Method n°1, how to inject your data in native charts?

  • Insert a chart with native tools of LibreOffice or MS Word in your document
  • Use traditional Carbone tags to create loops in chart's data to inject your JSON data
  • If necessary, use the special tag bind to tell Carbone that the value X in the chart must be replaced by the tag Y

  Method n°2, how to do advanced charts with Apache ECharts objects?

  • Insert a sample image in your template.

  • Place a tag in alt text , like a dynamic image : {d.chartObj:chart} with the formatter :chart. The formatter :chart is optional if the chartObj object contains the attribute "type" : "echarts@v5". In that case, Carbone automatically considers it is a chart object instead of a dynamic image.

  • chartObj must contains a compatible Echarts option. Here is an example to draw this chart:

    {
      "myChart" : {
        "type"    : "echarts@v5", // default
        "width"   : 600,          // default
        "height"  : 400,          // default
        "theme"   : null,         // default or object `theme` generated by https://echarts.apache.org/en/theme-builder.html
        "option"  : {
          "xAxis": {
            "type": "category",
            "data": ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
          },
          "yAxis": {
            "type": "value"
          },
          "series": [{
            "data": [150, 230, 224, 218, 135, 147, 260],
            "type": "line"
          }]
        }
      }
    }

  Currently, Carbone supports only "echarts@v5" but we may support newer versions and other libraries in the future. By default, Carbone considers "echarts@v5".

  Some charts have some translation: Locales supported : cs, de, en, es, fi, fr, it, ja, ko, pl, pt-br, ro, ru, si, th, zh

  Rendering charts with Apache Echarts is extremely powerful and works well if all these conditions are met

  • ECharts supports what you ask
  • The template supports the rendered SVG (docx/xslx/pptx does not support SVG images)
  • Your chart configuration does not need external dependencies (maps, js code, themes), which are not available in Carbone

  If you meet some limitation, please feel free to contact us on our chat to solve the issue.

• [EE] If options.isDebugActive = true. POST /render returns additional information in debug sub-object:

  {
    "renderId" : "file.pdf",
    "debug"    : {
      "markers" : ["{d.id}", "{d.tab[i].id}"] // all tags found in template
      "sample" : {         // EXPERIMENTAL
        "data"       : {}, // fake data generated from tags found in tags
        "complement" : {}  // fake complement generated from tags found in tags
      }
    }
  }

• Improve syntax error message:

  • when a tag tries to access an array and a object in the same time
  • when there is a missing [i] tag fo one [i+1] tag
  • when Carbone cannot find the section to repeat
  • when there is a dot . before []

• Fix crash when repetition does not contain XML tags. For example: <w:t>{d[i].id}, {d[i+1].id}</w:t>

• Fix crash when the section i+1 is duplicated like the i-th section with nested repetition and other tags

• Fix crash when repetition uses direct access of sub-arrays {d.test.others[i].wheels[0].size} {d.test.others[i+1].wheels[0].size}

• [EE] On-Premise Embedded Studio has new features and fixes:

  • [EXPERIMENTAL]: sample Data and Complement are automatically generated using tags found in template if these field contain empty objects
  • export to other formats than PDF
  • change report language
  • fix firefox template upload
  • fix memory leak
  • Now it works on Safari, without hot-reloading of the template

• Formatters managements has been completely rewritten, to make it faster and more reliable. Here are acceptable syntax for formatters. For backward-compatibility: text containing single quotes are accepted if it does not contain a comma , before or after the single quote: anyFormatter(' text ,containing ' sin,gle ' quote ')

• Dynamic parameters passed in formatters with the dot . syntax has been improved:

  • Add the possibility to access sub arrays with positive integers: .sort[12].id
  • Add a syntax checker, which returns an understandable error when the syntax is not correct. Ex: .sort.[sd, .sor]t.[sd
  • Improved access performance by a factor of 10
  • Add the possibility to access array iterators of currently visited arrays. The number of dots equals the number of previous i. Example: In {d[i].cars[i].other.wheels[i].tire.subObject:add(.i):add(..i):add(...i)}
    • .i matches with the index of wheels[i]
    • ..i matches with the index of cars[i]
    • ...i matches with the index of d[i]

• [EE] ⚡️ New aggregator formatters : aggSum, aggAvg, aggMin, aggMax, aggCount

  Data

      { brand : 'Lu' , qty : 1  , sort : 1 },
      { brand : 'Fa' , qty : 4  , sort : 4 },
      { brand : 'Vi' , qty : 3  , sort : 3 },
      { brand : 'Fa' , qty : 2  , sort : 2 },
      { brand : 'To' , qty : 1  , sort : 1 },
      { brand : 'Vi' , qty : 10 , sort : 5 }

  Template => Result

  Global aggregation, without iterator
  

  • {d.cars[].qty:aggSum} => 21
  • {d.cars[].qty:aggAvg} => 3.5
  • {d.cars[].qty:aggMin} => 1
  • {d.cars[].qty:aggMax} => 10
  • {d.cars[].qty:aggCount} => 6

  Global aggregation, with array filters
  

  • {d.cars[sort>1].qty:aggSum} => 19
  • {d.cars[sort>1].qty:aggAvg} => 4.75
  • {d.cars[sort>1].qty:aggMin} => 2
  • {d.cars[sort>1].qty:aggMax} => 10
  • {d.cars[sort>1].qty:aggCount} => 4

  Global aggregation, with array filters, with other formatters
  

  • {d.cars[sort>1].qty:mul(.sort):aggSum:formatC} => 81.00 €
  • {d.cars[sort>1].qty:mul(.sort):aggAvg:formatC} => 13.50 €
  • {d.cars[sort>1].qty:mul(.sort):aggMin:formatC} => 1.00 €
  • {d.cars[sort>1].qty:mul(.sort):aggMax:formatC} => 50.00 €
  • {d.cars[sort>1].qty:mul(.sort):aggCount:formatC} => 6.00 €

- [EE] 🤩 Aggregators can also be used in a loop to compute sub-totals and cumulative totals (or "running totals"), with custom grouping clause (partition)
  - By default if not grouping clause is defined:
    - Sum by departments of all people's salary:
      - `{d.departments[i].people[].salary:aggSum}`
      - `{d.departments[i].people[].salary:aggSum(.i)}` (alternative)
    - Global sum of all departments, all people's salary
      - `{d.departments[i].people[i].salary:aggSum}`
      - `{d.departments[i].people[i].salary:aggSum(0)}` (alternative)
    - Cumulative total (or "running total") by departments of all people's salary:
      - `{d.departments[i].people[].salary:cumSum}`
    - Cumulative total (or "running total") of all departments, all people's salary
      - `{d.departments[i].people[i].salary:cumSum}`

  - You can change the partition with dynamic parameters like that
    - Sum by people by age, regardless of departments
      - `{d.departments[i].people[i].salary:aggSum(.age)}`
    - Sum by people by age and gender, regardless of departments
      - `{d.departments[i].people[i].salary:aggSum(.age, .gender)}`

v3.7.2

• Release June 13th 2024
• Update dependencies for security
• Improve Libreoffice detection on Arch Linux (Thank you @jan-san)
• Fix: removed the possibility of prototype pollution in formatters

v3.7.1

• Release December 19th 2023
• [EE] Fix crash when dynamic image URL contains forbidden characters

v3.7.0

• Release June 15th 2023
• [EE] All APIs returns "400 Bad request" instead of "404 Not Found" when idTemplate is not valid
• [EE] Fixed HEAD /render/:renderId : The request don't delete the generated document anymore.
• [EE] Fixed GET /template/:templateId : If the template doesn't exist, the statusCode 404 is returned instead of 400.
• [EE] Force download of the rendered report when the query parameter ?download=true is set, such as: POST /render/:renderId?download=true
• [EE] On-Premise: New option to enable more security controls, following ANSSI (France) and BSI (Germany) recommendations. Contact us for more information.
• [EE] Updated DELETE /template/:templateId: If there isn't a template, Carbone does not remove any template but will still respond that the command was successful with a code 200.

v3.6.0

• Release April 17th 2023
• Support vector barcodes with the new svg option to improve print quality: {d.number:barcode(qrcode, svg:true)}

v3.5.6 [EE]

• Release February 16th 2023
• [EE] BREAKING CHANGE in Carbone On-Premise plugins: the callback of getPublicKey must take two arguments (err, publicKeys). If the first argument contains an error, the bad token is kept in quarantine area for 60s instead of infinitely.

v3.5.6

• Release June 12th 2024
• Fix: removed the possibility of prototype pollution in formatters. This can only occur if the parent NodeJS application has the same security issue. CVSS:3.0/AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H.
• Update some dependencies

v3.5.5

• Release December 7th 2022
• Add user-agent Carbone when webhook is called.

v3.5.4

• Release June 15th 2022
• [EE] Do not return an error when DEL /template is called and the template is already deleted on local storage. It may be already deleted by the plugin.

v3.5.3

• Release May 25th 2022
• [EE] Accept convCRLF before :html formatter to convert to <br>

v3.5.2

• Release May 6th 2022
• [EE] Add file verification on template upload

v3.5.1

• Release May 4th 2022
• Rollback fix in v3.4.9

v3.5.0

• Release May 4st 2022
• Formatters add(), mul(), sub() and div() accept simple mathematical expressions inside parenthesis.
  • Example: {d.val:add(.otherQty + .vat * .price - 10 / 2)
  • Only mathematical operators +, *, -, / are allowed, without parenthesis
  • Multiplication and division operators (*, /) has higher precedence than the addition/substration operator (+, -) and thus will be evaluated first.

v3.4.9

• Release April 27st 2022
• Fix crash with very complex JSON map

v3.4.8

• Release March 15st 2022
• [EE] Fix: avoid crash when a tag is used on a shape instead of a sample image (v3.2.2-1)
• [EE] Fix graceful exit on SIGTERM, keep the converter alive to finish remaining renders!
  • As soon as Carbone has finished all renders, it exits after 15 seconds instead of 10 seconds
• [EE] Fix DOCX documents that are including dynamic images and static charts

v3.4.7

• Release March 1st 2022
• [EE] Carbone-EE On-Premise accepts to read the license from environment variable CARBONE_EE_LICENSE, or --license CLI options

v3.4.6

• Release February 18th 2022
• [EE] The HTTP server starts as soon as possible, before LibreOffice.
• [EE] Gracefully exits on SIGTERM. When the signal is received
  • GET /status returns 503. The reverse-proxy should stop sending new requests
  • As soon as Carbone has finished all renders, it exits after 10 seconds
  • Whatever happens, it exists after 300 seconds.
• [EE] Add barcode options: pass to the formatter :barcode options as second argument. It should keep the format "key:value", such as: {d.number:barcode(qrcode, width:300, height:100, includetext:false)}. Options available:
  • width Width a number as millimeters. Example: {d.number:barcode(qrcode, width:300)}
  • height Height a number as millimeters. Example: {d.number:barcode(qrcode, height:10)}
  • scale Quality of the barcode as a number between 1 to 10. The default value is 3. Example: {d.number:barcode(qrcode, scale:3)}
  • includetext Show the text and takes a Boolean as value, true or false. Example: {d.number:barcode(qrcode, includetext:false)}.
  • textsize Number to change the size of the text. Example: {d.number:barcode(qrcode, textsize:20)}
  • textxalign Change the alignment of the text horizontally. Takes only 4 values: left, center, right, or justify. The default value is center. Example: {d.number:barcode(qrcode, textxalign:right)}
  • textyalign Change the alignment of the text vertically. Takes only 3 values: below, center, or above. The default value is below. Example: {d.number:barcode(qrcode, textyalign:above)}.
  • rotate Rotate the barcode and the text. Takes only 4 values (case insensitive): N for not rotated, R for 90 degree right rotation, L for 90 degree left rotation, I of 180 degree rotation. Example: {d.number:barcode(qrcode, rotate:R)} or {d.number:barcode(qrcode, rotate:l)}.
  • barcolor Color of bars as hexadecimal #RRGGBB. Example: {d.number:barcode(qrcode, barcolor:#1FDE25)}. Note: 6 characters required, and case insensitive.
  • textcolor Color of the text as hexadecimal #RRGGBB. Example: {d.number:barcode(qrcode, textcolor:#1FDE25)}. Note: 6 characters required, and case insensitive.
  • backgroundcolor Color of the background as hexadecimal #RRGGBB. Example: {d.number:barcode(qrcode, backgroundcolor:#1FDE25)}. Note: 6 characters required, and case insensitive.
  • eclevel Specify the error correction level: L for Low, M for Medium (default), Q for Quality and H for High. Option ONLY FOR QRCODES, Micro QR Code, GS1 QR Code, HIBC QR Code, or Swiss QR Code.
• [EE] Dynamic HTML improvement - the following styles applied on the :html formatter are kept on the generated document: Right-to-left text, and text/background colors.
• [EE] Dynamic HTML Fix - When creating a ordered or unordered list, the font-size, and font-family applied on the template are now kept in the generated document.

v3.4.5

• Release February 9th 2022
• Now the parameter converterFactoryTimeout updates also the HTTP socket timeout accordingly

v3.4.4

• Release February 7th 2022
• [EE] Experimental: increase the limit to 200 repetitions maximum when using the repetition feature {d[i+1*qty]

v3.4.3

• Release January 31th 2022
• [EE] Bump dependencies to latest version
• [EE] Add options.isDebugActive. If true, POST /render returns additional information in debug sub-object:

  {
    "renderId" : "file.pdf",
    "debug"    : {
      "markers" : ["{d.id}", "{d.tab[i].id}"] // all tags found in template
    }
  }

v3.4.2

• Release January 10th 2022
• [EE] Fix do not crash if content passed to :html formatter is not a string

v3.4.1

• Release December 9th 2021
• Accepts "OpenDocument Text Flat XML" (.fodt) template files
• Includes v3.3.3: fix timezone conversion with latest IANA database to manage correctly Daylight Saving Time
• [EE] Fix dynamic colors for DOCX cells background

v3.4.0

• Release November 17th 2021
• Remove compatibility with NodeJS 10.x. V8 uses timsort since NodeJS 11. So we can remove timsort dependency. NodeJS 12+ required.
• Bump DayJS to 1.10.7 and debug to 4.3.2
• Improve thread management of LibreOffice on Linux:
  • The system, which auto-restarts LibreOffice when it crashes or if there is a conversion timeout, could hang indefinitely on Linux. On the Enterprise Edition, the global watchdog system was able to fix this bad behavior but it was slow. Now, the LibreOffice is correctly killed. No zombie processes remaining.
  • Avoid launching the parent process "oosplash" of LibreOffice
  • Improve auto-restart mechanism
  • Add debug logs
  • All tests passed on Linux 😅
• [EE] Bump all dependencies
• [EE] Fix random image display in LibreOffice documents. Sometimes, LibreOffice hides one image when two or more images share the same name.

   Now, Carbone generates a unique name for each image with the format "carbone-image-<counter>".

• [EE] Experimental: Deffered rendering with a webhook, it is dedicated to render huge reports:
  1. Render a document as usual with the request POST /render/:templateID with the JSON dataset into the body request AND you have to insert the header carbone-webhook-url as a callback URL. It is an endpoint of your server listening when a document is rendered.
  2. Carbone will generate your document and it will notify your webhook with a renderID
  3. Retrieve the generated document with a GET /render/:renderID and voilà!
• [EE] Experimental: add the possibility to duplicate rows using an attribute of an object Data:

    [
      { "id" : "A", "qty" : 2 },
      { "id" : "B", "qty" : 3 },
      { "id" : "C", "qty" : 0 },
      { "id" : "D", "qty" : 1 }
    ]

  Template: {d[i].id} - {d[i+1*qty].id} Result: A - A - B - B - B - D
• [EE] ⚡️ Carbone supports 107 barcodes in DOCX/ODT/XLSX/ODS templates:
  • Barcodes are inserted as a dynamic image to support more types
  • In your template, barcodes tags must be inserted inside the title or description field of a temporary image, and then it must be followed with the barcode formatter, such as {d.value:barcode(type)}.
  • You must pass one of the following types to the :barcode formatter as a first argument: ean5, ean2, ean13, ean8, upca, upce, isbn, ismn, issn, code128, gs1-128, ean14, sscc18, code39, code39ext, code32, pzn, code93, code93ext, interleaved2of5, itf14, identcode, leitcode, databaromni, databarstacked, databarstackedomni, databartruncated, databarlimited, databarexpanded, databarexpandedstacked, gs1northamericancoupon, pharmacode, pharmacode2, code2of5, industrial2of5, iata2of5, matrix2of5, coop2of5, datalogic2of5, code11, bc412, rationalizedCodabar, onecode, postnet, planet, royalmail, auspost, kix, japanpost, msi, plessey, telepen, telepennumeric, posicode, codablockf, code16k, code49, channelcode, flattermarken, raw, daft, symbol, pdf417, pdf417compact, micropdf417, datamatrix, datamatrixrectangular, datamatrixrectangularextension, mailmark, qrcode, swissqrcode, microqrcode, rectangularmicroqrcode, maxicode, azteccode, azteccodecompact, aztecrune, codeone, hanxin, dotcode, ultracode, gs1-cc, ean13composite, ean8composite, upcacomposite, upcecomposite, databaromnicomposite, databarstackedcomposite, databarstackedomnicomposite, databartruncatedcomposite, databarlimitedcomposite, databarexpandedcomposite, databarexpandedstackedcomposite, gs1-128composite, gs1datamatrix, gs1datamatrixrectangular, gs1qrcode, gs1dotcode, hibccode39, hibccode128, hibcdatamatrix, hibcdatamatrixrectangular, hibcpdf417, hibcmicropdf417, hibcqrcode, hibccodablockf, hibcazteccode
  • The previous system, which uses a special font, is still available but is limited to ean8, ean13, ean128, code39.

v3.3.2

• Release October 11th 2021
• [EE] Dynamic Image fix: image types verification support uppercase and lower case formats

v3.3.1

• Release September 9th 2021
• [EE] New Carbone On-premise: Pass the option "maxDataSize" to change the maximum JSON data size when rendering a report. The value must be bytes. The default value is 60MB.

v3.3.3

• Release November 26th 2021
• Fix timezone conversion with latest IANA database to manage correctly Daylight Saving Time 2021-11-18T08:05+0000 -> Europe/London -> Thursday, November 18, 2021 8:05 AM

v3.3.0

• Release August 30th 2021
• [EE] HTML Formatter:
  • Fix: The HTML content is rendered without adding an empty line above it.
  • Fix: The HTML content and static content are rendered in the expected order.
  • New: If static content and Carbone tags are mixed with an HTML formatter in the same paragraph, the html is isolated into a new paragraph and each element are seperated above or below. For example, the following template on a text editor <paragraph>A rocket is made of {d.data:html} {d.details}, this is cool!</paragraph> will be transform into 3 paragraphs on the generated report <paragraph>A rocket is made of </paragraph><paragraph>{d.data:html}</paragraph><paragraph> {d.details}, this is cool!</paragraph>.
  • Improved HTML rendering stability when it is mixed with lists, tables, and images.
• [EE] Dynamic Checkbox are supported only for ODT file. A tag should be inserted into the checkbox property "name" and it is used to set the value of the checkbox on the generated report. The checkbox is ticked (checked) when the value is a Boolean with the value "true", a non empty string, a non empty array or a non empty object. If the exported file type is a PDF, the checkbox can be edited on the generated document. An ODT document created from MS Word that include checkboxes does not work. It is also not possible to create a list of checkboxes with the expressions [i] / [i+1].
• Accept null for the attribute complement in options

v3.2.7

• Release July 21th 2021
• Fix corrupted document when accessing a sub-object in an array {d.surrounding[i].subArray[0].subObject.id}, within a surrounding loop

v3.2.6

• Release June 15th 2021
• [EE] Fix the generation of ODP document that includes table lists.

v3.2.5

• Release June 10th 2021
• [EE] Accept URLs with weird Content-Type such as image/png; charset=utf-8 for dynamic image replacement

v3.2.4

• Release May 25th 2021
• [EE] Add the possibility to upload templates in base64. The content-type must be application/json and the template must be sent in base64 in the body { "template" : "pure base64 or data-URI scheme in base64"}
• [EE] Accepts loops with dynamic image replacement across slides/pages in ODP templates

v3.2.3

• Release May 21th 2021
• Accepts letter W to get the week number in formatD formatter

v3.2.2-1

• Release March 11th 2022
• Fix: avoid crash when a tag is used on a shape instead of a sample image

v3.2.2

• Release May 10th 2021

• Fix broken Excel files. It removes the alert which appears sometime when opening the file with MS Excel.

• Update DayJS dependency from 1.9.6 to 1.10.4.

• Fix date formatters (formatD...) and number formatters (formatN...) when the country code is used in options.lang. It accepts lower case or upper case for the locale. Example:

  Data:

    {
      date  : '20140131 23:45:00',
      price : 1000.1234
    }

  Template: {d.date:formatD(dddd)} {d.price:formatN()}

  Result:

  • de-DE => Friday 1.000,123 (before) Freitag 1.000,123 (after)
  • fr-fr => Friday 1 000,123 (before) vendredi 1 000,123 (after)

• Fix number formatting for these locales: nl-be, am, ar, ar-dz, ar-bh, ar-eg, ar-iq, ar-jo, ar-kw, ar-lb, ar-ly, ar-ma, ar-om, ar-qa, ar-sa, ar-sy, ar-tn, ar-ae, ar-ye, az-az, bn, bs, zh-cn, zh-hk, zh-mo, zh-sg, zh-tw, hr, da, en-au, en-bz, en-ca, en-cb, en-in, en-ie, en-jm, en-nz, en-ph, en-tt, fa, fr-lu, de-li, de-lu, de-ch, el, he, hi, id, it-it, it-ch, ja, kn, ko, ms-bn, ms-my, ml, mr, ro-mo, ro, ru-mo, sr-sp, sl, es-ar, es-bo, es-cl, es-co, es-do, es-ec, es-sv, es-gt, es-hn, es-mx, es-ni, es-pa, es-py, es-pe, es-pr, es-es, es-uy, es-ve, sw, ta, te, th, tr, uz-uz, vi, de, es, it

  Example:

  • de => Freitag 1,000.123 (before) Freitag 1.000,123 (after)

v3.2.1

• Release May 4th 2021
• Fix locale de-de
• [EE] Fix dynamic HTML: null or undefined values return an empty string instead of an error.

v3.2.0

• Release April 13th 2021
• [EE] Fix dynamic image resize when using the :imageFit formatter with the contain property.
• [EE] Add the option :imageFit(fillWidth) to fill the full width of the template image while keeping aspect ratio of the inserted image. Before fixing the bug with contain property, it was more or less the default behaviour of Carbone before this version. So, fillWidth becomes the default option to avoid changing the style of existing reports.
• [EE] Improve performance to download images. Building a report with a lot of dynamic images is almost 5 times faster than before.
• [EE] When the dynamic image cannot be inserted (fetch failed, image type not supported),
  • the aspect ratio of the replacement image keep the aspect ratio of the template
  • this replacement image more beautiful (vectorial) and does not contain any text for internationalisation
• [EE] Fix corrupted XLSX files when inserting new type of image which were not previously present

v3.1.7

• Release April 12th 2021
• [EE] Fix Dynamic HTML: paragraph and break lines inside nested lists or anchor tags were creating corrupted DOCX/ODT documents.

v3.1.6

• Release April 12th 2021
• [EE] New formatter :defaultURL(): if a dynamic hyperlink or a HTML anchor tag is injected into a report and the URL verification fails, the formatter is used to replace the default error URL. Example to use it with and HTML formatter: {d.content:defaultURL(https:url.of.your.choice):html}. The :defaultURL should be placed before the :html formatter.
• [EE] Fix: return 404 error when the template does not exist on rendering

v3.1.5

• Release April 8th 2021
• [EE] Fix hyperlinks verification which could lead to crash.

v3.1.4

• Release April 1st 2021
• [EE] Improve :html formatter stability when using special characters such as "'<>& for ODT and DOCX reports.

v3.1.3

• Release March 29th 2021
• Fix: Do not break documents if the i+1 row contains some tags coming from parent object or condition blocks (rare)
• [EE] if a font family and font size is applied to an HTML formatter {d.content:html}, the font & size will be applied to the whole rendered HTML
• [EE] return an error message when image anchor is not correct in the template

v3.1.2

• Release March 4rd 2021
• Fix: v3.1.0 introduced a backward compatibility issue with reports made with v1/v2. Now, filter with boolean works like this (same behavior as numbers)
  • data => template => condition result in array
  • data.myBoolean = true => d.array[i, myBoolean=true] => true
  • data.myBoolean = "true" => d.array[i, myBoolean=true] => true
  • data.myBoolean = "false" => d.array[i, myBoolean=true] => false
  • data.myBoolean = false => d.array[i, myBoolean=true] => false
  • data.myBoolean = "true" => d.array[i, myBoolean='true'] => true
  • data.myBoolean = true => d.array[i, myBoolean='true'] => false

v3.1.1

• Release March 4rd 2021
• [EE] Fix: remove html entities not supported by XML format when using :html formatter

v3.1.0

• Release March 3rd 2021
• Accepts boolean in array filters d.array[i, myBoolean=true]
• [EE] Improve hyperlinks validation
• [EE] Improved HTML conversion with formatter :html for DOCX / ODT templates:
  • Support ol, ul, p, ul, ol, li and a tags
  • Fixed spacing management between list / paragraph / multi elements.
  • All HTML entities are supported (Full List: https://www.w3schools.com/charsets/ref_html_entities_4.asp)
  • Improved break-lines support
  • Fixed hyperlinks when exporting DOCX to PDF
  • Possible to include hyperlinks, break-lines, and style tags into any level of lists
  • Improved overall stability

v3.0.4

• [EE] Restore old convert formatter for some clients

v3.0.3

• [EE] Fix: add the "https://" protocol if it is missing for dynamic hyperlinks

v3.0.2

• [EE] Fix: accepts hyperlinks with "&", add the "https://" protocol if it missing, add URL validation. If the URL is invalid, it is replaced by a valid URL refering to the carbone documentation.
• [EE] Fix: do not crash if hyperlink is undefined
• [EE] Fix: support random way of managing hyperlinks in MS Word
• [EE] Fix: set Content-Type when downloading the report

v3.0.1

• Fix: aliases beginning with same prefix names are properly rendered in the generated reports instead of not being skip.
• [EE] Accepts a new local filename in the On-Premise plugin writeTemplate(err, newFilename)
• [EE] Fix: license detection in Docker
• [EE] Fix dynamic hyperlinks for DOCX reports
• Improve documentation

v3.0.0

• 👋🏻 NOTE: This version contains breaking changes of undocumented features. So if you use only documented features so far, you should not be concerned by these breaking changes.

• ⚡️ Manage timezone + new date formatters + switch from MomentJS to DayJS

  • If not defined by you in options.complement, {c.now} returns the current date in UTC.

  • [BREAKING CHANGE]: remove old date formatter which were not documented: format, parse, addDays and convert. You should use formatD instead and new formatters below. They were very old formatters, the chance you use them is low because you had to look into the source code to know their existence.

  • New formatters:

    • addD(amount, unit [, patternIn]) : add days, month to a date. formatD can be used after without specifying patternIn
    • subD(amount, unit [, patternIn]) : subtract days, month to a date. formatD can be used after without specifying patternIn
    • startOfD(amount, unit [, patternIn]) : Set a date to the start of a unit of time. formatD can be used after without specifying patternIn
    • endOfD(amount, unit [, patternIn]) : Set a date to the end of a unit of time. formatD can be used after without specifying patternIn

  • [BREAKING CHANGE]: We try to stay as close as possible as the previous parsing algorithm. But 1997-12-17 07:37:16-08 is not accepted anymore without specifying an input pattern or writing 1997-12-17 07:37:16-08:00

  • Accept a new options to set the timezone of formatD. Examples:

    Data

      {
        date     : '2020-11-28T21:54:00.000Z',  // ISO 8601 UTC
        dateWTZ  : '2020-11-28T21:54:00',       // without timezone
        dateTZ   : '2020-11-28T21:54:00-04:00', // with America/New_York timezone offset
        dateUnix : 1606600440                   // UNIX timestamp in seconds UTC of 2020-11-28T21:54:00.000Z
      }

    Template => Result

    if options.timezone = 'Europe/Paris' (default)

    • {d.date:formatD(LLLL)} => Saturday, November 28, 2020 10:54 PM
    • {d.dateWTZ:formatD(LLLL)} => Saturday, November 28, 2020 9:54 PM
    • {d.dateTZ:formatD(LLLL)} => Sunday , November 29, 2020 2:54 AM
    • {d.dateUnix:formatD(LLLL, X)} => Saturday, November 28, 2020 10:54 PM

    if options.timezone = 'America/New_York'

    • {d.date:formatD(LLLL)} => Saturday, November 28, 2020 4:54 PM
    • {d.dateWTZ:formatD(LLLL)} => Saturday, November 28, 2020 3:54 PM
    • {d.dateTZ:formatD(LLLL)} => Saturday, November 28, 2020 8:54 PM
    • {d.dateUnix:formatD(LLLL, X)} => Saturday, November 28, 2020 4:54 PM

    List of timezone: https://en.wikipedia.org/wiki/List_of_tz_database_time_zones

• Fix: if a path does not exist inside a formatter argument, it returns an empty string instead of the error "[[C_ERROR]] attribute_name not defined". It fixes some weird behaviour with ifEM formatters

• [EE] Feature: cells colors on ODT/DOCX report can be changed dynamically with the "bindColor" tag.

• [EE] ODT Improvement: the "bindColor" tag will not remove other styles than colors.

• Accepts to convert the first page of docx or odt templates into a JPEG file with converTo : 'jpg'

• Improve HTML type detection. Accepts html5 without doctype.

• [EE] Fix Carbone tag inside ODT text box

• Adding padl and padr string formatter.

• Fix doc issue on carbone website

• Accepts Adobe Indesign IDML file as a template

• [EE] Dynamic hyperlinks: it is possible to insert hyperlinks into elements (text, image, list, tables, ...). Right click an element, select "hyperlinks", insert the tag and validate. It is working with ODS, ODT, and DOCX reports. The compatibility is limited for XLSX documents: It is not possible to create a list of hyperlinks and the tag should not be written with curly braces, example: a typical {d.url} should be only d.url. If http:// appears before d.url, it is also valid.

• Improve the parsing processing by moving the function "removeXMLInsideMarkers" before the building stage.

• Support officially to embed translations tags inside other tags: {d.id:ifEq(2):show( {t(Tuesday)} ) }

• Performance: reduce disk IO when converting document

• Performance: deactivate image compression by default to speed up PDF conversion

• [BREAKING CHANGE]: remove the possibility to use convertTo.formatOptionsRaw for CSV export. This feature was not documented and can lead to security issues. Use convertTo.formatOptions instead.

• new paramater in Carbone.set

  • renderPath : Carbone.set can changes the default path where rendered files are temporary saved.

                 By default, it creates the directory `carbone_render` in Operating System temp directory.
                 It creates the path automatically

• new paramater in Carbone.render

  • renderPrefix : If defined in options object. Carbone.render returns a file path instead of a buffer, and it adds this prefix in the rendered filename

                 The generated filename contains three parts:
                   - the prefix
                   - a secure Pseudo-Random part of 22 characters
                   - the report name, encoded in specific base64 to generate safe POSIX compatible filename on disk
                 `/renderpath/<prefix><22-random-chars><encodedReportName.extension>`
                 This filename can be decoded with the function `Carbone.decodeOuputFilename(pathOrFilename)`.
                 It is the user responsability to delete the file or not.

• New function decodeOuputFilename(): when renderPrefix is used, the returned filename can be parsed with this function.

                  It decodes filename an returns an object with two parameters
                  ```js
                  {
                    extension  : 'pdf',                // output file extension
                    reportName : 'decoded report name' // reportName
                  }
                  ```

• [BREAKING CHANGE]: Carbone.convert function signature has changed. Now, it accepts the same options as Carbone.renders: You must use Carbone.convert(fileBuffer, options, callback) instead of Carbone.convert(fileBuffer, convertTo, options, callback)

• [EE] Returns errors when Carbone cannot dynamically replaces images (ex. images with absolute positions) instead of generating a bad report

• [EE] Fix some PDF options. Integer values were not taken into account.

• [EE] Add the possibilty to generate JPG or PNG of the first page of a document Available options for jpg/png:

      convertTo : {
        formatName    : 'jpg', // or png
        formatOptions : {
          Quality     : 90,  // [JPG ONLY] From 1 (low quality, high compression, low size) to 100 (high quality, low compression, high size
          PixelWidth  : 100, // Image width as pixels
          PixelHeight : 100, // Image height as pixels
          Compression : 90,  // [PNG ONLY] From 0 (compression disabled) to 9 (high compression, low size)
          Interlaced  : 0,   // [PNG ONLY] 0 not interlaced, 1 enterlaced (higher size)
          ColorMode   : 0    // 0 Colors, 1 Greyscale
        }
      }

• 🎁 [EE] DOCX/ODT New feature: Support HTML rich content, by adding the :html formatter, it is possible to render the following HTML tag: <br>/<b>/<strong>/<i>/<em>/<u>/<s>/<del>. Unsupported tags and tags attributes are skipped and not rendered. HTML entities are accepted.

• [EE] New feature: Dynamic pictures are supported on ODG and ODP files. It is not possible to create loops.

• [EE] New Carbone Render On-Premise

  • Carbone can be safely deployed on your own servers. Contact us for further information

v2.1.1

• Release September 23rd 2020
• Fixes arrayJoin( ):convCRLF. Now it works in carbone v2.x.x like in v1.x.x.
• Removes 'zipfile' dev dependency. Tests use unzip from the system instead.
• 8.1.3 mocha upgrade
• [EE] Fixes crash when images field in data contain an object instead of a string

v2.1.0

• Release September 1st, 2020
• Performance: huge gain from x11 to x30 for the compression of reports. Now, some huge reports takes 0.1s to render instead of 4s. It reduces also the blocking of Node's event loop.
• New rendering option hardRefresh: The content of the report is refreshed at the end of the Carbone process. For example, it can be used to refresh the table of content or update calculations. The option convertTo has to be defined.
• [EE] fix dynamic images in DOCX files that were making the report invalid in Word
• [EE] fix dynamic images in header & footers of docx templates
• [EE] Injecting dynamic colors received a lot of improvements and stability:
  • ODT, DOCX, and ODS reports are fully supported. XLSX can't be supported by design for now.
  • Text and background colors in footers and headers are supported for ODT and DOCX templates.
  • Better error management, it throws errors when:
    • bindColor is not correctly formatted
    • 2 bindColor tags try to edit the same color
    • the background color format on DOCX documents is different than "color"
    • the color format does not exist
    • 2 different lists of colors are used to edit the same element

v2.0.2

• Release August 10th, 2020
• Fix locales de-ch and pt-br
• Fix direct access in a nested array {d.array[i].nestedArray[i=0].id}
• [EE] : fix server exit on "Ctrl+C"

v2.0.1

• Release July 8th, 2020
• Add regression tests
• [EE] : fix crash when an array was printed directly without formatters {d.myArray}

v2.0.0

• Release June 28th, 2020

• 🚀 Accepts dynamic variables in all formatters!

  Carbone passes data to formatters if parameters start with a dot . and is not surrounded by quotes. Here is an example:

  Data

      {
        id : 10,
        qtyA : 20,
        subObject : {
          qtyB : 5,
          qtyC : 3
        },
        subArray : [{
          id : 1000,
          qtyE : 3
        }]
      }

  Template => Result

  do mathematical operations:
  {d.subObject.qtyB:add(.qtyC)} => 8 (5+3)

  read parent attributes if you use two dots, grandparents if you use three dots, etc...
  {d.subObject.qtyB:add(.qtyC):add(..qtyA)} => 28 (5+3+20)

  read parent objects and their children attributes (no limit in depth)
  {d.subArray[i].qtyE:add(..subObject.qtyC) => 6 (3+3)

  It returns an error if the attribute does not exist
  {d.subArray[i].qtyE:add(..badAttr) => [[C_ERROR]] badAttr not defined

  You cannot access arrays
  {d.subObject.qtyB:add(..subArray[0].qtyE)} => [[C_ERROR]] subArray[0] not defined

• ⚡️ New conditional formatters, and a new IF-block system to hide/show a part of the document

  • ifEQ (value) : Matches values that are equal to a specified value, it replaces ifEqual
  • ifNE (value) : Matches all values that are not equal to a specified value
  • ifGT (value) : Matches values that are greater than a specified value.
  • ifGTE (value) : Matches values that are greater than or equal to a specified value.
  • ifLT (value) : Matches values that are less than a specified value.
  • ifLTE (value) : Matches values that are less than or equal to a specified value.
  • ifIN (value) : Matches any of the values specified in an array or string, it replaces ifContain
  • ifNIN (value) : Matches none of the values specified in an array or string
  • ifEM (value) : Matches empty values, string, arrays or objects, it replaces ifEmpty
  • ifNEM (value) : Matches not empty values, string, arrays or objects
  • and (value) : AND operator between two consecutive conditional formatters
  • or (value) : (default) OR operator between two consecutive conditional formatters
  • hideBegin and hideEnd : hide text block between hideBegin and hideEnd if condition is true
  • showBegin and showEnd : show a text block between showBegin and showEnd if condition is true
  • show (message) : print a message if condition is true
  • elseShow (message) : print a message if condition is false
  • len() : returns the length of a string or array.

  No formatters can be chained after hideBegin, hideEnd, showBegin, showEnd.

  These new formatters replace the old ones ifEqual, ifEmpty and ifContain. We keep these old formatters for backwards compatibility. You should avoid using them in new templates.

  Data

      {
        id : 10,
        qtyA : 20
      }

  Template => Result

  print simple message according to the result of a condition
  {d.id:ifEQ(10):show('hey')} => hey

  print simple message according to the result of multiple conditions, combining with dynamic variables! 🤩
  {d.id:ifEQ(10):and(.qtyA):ifEQ(20):show('hey'):elseShow('hide')} => hey

  hide or show a block of text in the document ⚡️
  {d.id:ifEQ(10):showBegin} block of text {d.id:showEnd} => block of text
  {d.id:ifEQ(12):showBegin} block of text {d.id:showEnd} =>

  A smart and generic algorithm detects automatically pattern (paragraph, bullet-list, etc) to remove in document even if the conditional block is not placed correctly in XML. For example: BEGIN<p> blabla END</p> becomes BEGIN<p> blabla </p>END. It improves the final result by removing empty spaces in document.

• ☀️ Accepts to iterate on attributes of objects as is if it was an array

  {
    myObject : {
      paul : '10',
      jack : '20',
      bob  : '30'
    }
  }

  In the report:

    {d.myObject[i].att} {d.myObject[i].val}
    {d.myObject[i+1].att} {d.myObject[i+1].val}

  • use .att to print the attribute
  • use .val to print the value

  You can even access nested objects and nested arrays inside val: {d.myObject[i].val.myArray[i].id}

• Fix: avoid crashing when a static image was inserted in a loop in MS Word templates

• Update totals in ODS and XSLX files

• formatC supports many crypto currencies

• Fix: Accepts comma in formatter parameters such as {d.sentenceExist:ifEqual(1, 'Some sentence, some more sentences.')}

• Fix: nested array in XML (but not in JSON) was not printed correctly

    {d.countries[i].name}
      {d.movies[i].subObject.name}
      {d.movies[i+1].subObject.name}
    {d.countries[i+1].name}

• Fix: avoid crashing when a sub-object is null or undefined in data

• Fix: avoid crashing when the parent object of an array is null or undefined in data

• Eslint code + add eslint tools

• Fix: accepts dashes characters in JSON data. Before, Carbone crashes when using {d.my-att-with-dash}

• Fix: avoid crashing when a XLSX template contains charts

• Beta: supports dynamic charts rendering in XLSX if these conditions are met:

  • first, draw a chart in MS Excel and replace your data with Carbone tags
  • datas of the chart should be placed at the top-left corner of the spreadsheet
  • all numbers are formatted with formatN() formatter

• Fix: accepts white-space in array filters with simple quote and double quotes Example: {d.cars[i, type='Tesla car'].name}

       `{d.cars[i, type="Tesla car"].name}`

• Fix LibreOffice detection on Windows

• Remove compatibility with older NodeJS versions (lower than 10.15.0)

• Upgrade some dependencies (moment, debug, yauzl) and remove useless ones (should)

• Accepts non-alphanumeric characters in variables names, values, ... For example, {d.i💎d} is allowed

• Fix many security issues and reduce memory consumption

• Fix crash when tags are next to each over {d.id}{d.other} in many situations:

  • with or without conditional blocks
  • with or without loops

• Fix crash when some documents like DOCX contain images in repetition section

• Accept direct access in arrays such as {d.myArray[2].val} instead of {d.myArray[i=2].val}

• Fix crash when two consecutive arrays, nested in object, were used

• Remove useless soft-page-break in ODT documents as suggested by the OpenDocument specification

• LibreOffice 6+ has a memory leak. So Carbone automatically restarts LibreOffice after a certain amount of document conversion. The number of conversions depends on new parameters factoryMemoryFileSize and factoryMemoryThreshold

• Add conversion timeout parameter converterFactoryTimeout (60s by default). It kills LibreOffice if the conversion is too long and returns an error

• Remove deprecated NodeJS "new Buffer"

• Fix: avoid crashing if a object/array is null or undefined. Print empty text instead.

• Fix: variables, which begin by the same characters, were not detected correctly since NodeJS 11

• [EE] Image processing completely rewritten

• [EE] Dynamic images improvements: it is possible to insert images into ODT, ODS, XLSX and DOCX by passing a public URL or a Data URLs. For the 2 solutions, you have to insert a temporary picture in your template and write the tag as an alternative text. Finally, during rendering, Carbone replaces the temporary picture by the correct picture provided by the tag.

  The place to insert the tag on the temporary picture may change depends on the file format:

  • ODS file: set the tag on the image title
  • ODT file: set the tag on the image alternative text
  • DOCX file: set the tag either on the image title, image description, or alternative text
  • XLSX file: set the tag either on the image title, image description, or alternative text

  The accepted images type are: png, jpeg/jpg, gif, svg

  If an error occurs for some reason (fetch failed, image type not supported), a replacement image is used with the message "invalid image".

• [EE] dynamic images: new formatter :imageFit() only available for DOCX and ODT files. It sets how the image should be resized to fit its container. An argument has to be passed to the formatter: contain or fill. If the formatter is not defined, the image is resized as contain by default.

  • contain: The replaced image is scaled to maintain its aspect ratio while fitting within the element’s content-box (the temporary image).
  • fill: The replaced image is sized to fill the element’s content-box (the temporary image). The entire image will fill the box of the previous image. If the object's aspect ratio does not match the aspect ratio of its box, then the object will be stretched to fit.

  example: {d.myImage:imageFit(contain)} or {d.myImage:imageFit(fill)}

• [EE] Added Libreoffice export filters for PDF rendering. To apply filters to a PDF, it is possible to assign an object to convertTo with formatName:'pdf' and formatOptions an object with the specified filters. The filter list is available on the documentation. Here is an example of a PDF with a password and a watermark:

    options = {
      convertTo: {
        formatName: 'pdf',
        formatOptions: {
          EncryptFile          : true,
          DocumentOpenPassword : 'QWERTY1234',
          Watermark            : 'Watermark Carbone'
        }
      }
    }

v1.2.1

• Release June 11, 2019
• Fix arrayMap() if used with an array of strings or integer

v1.2.0

• Release March 13, 2019
• Add new formatters
  • convCurr(targetCurrency, sourceCurrency) to convert from one currency to another
  • formatN() format number according to the locale (lang). Examples:
    • old toFixed(2):toFR can be replaced by formatN(2)
  • formatC() format currency according to the locale and the currency
    • old toFixed(2)} {t(currency)} can be replaced by formatC(2)
  • formatD() format date according to the locale. Same as convDate, but consider parameters are swapped for consistency with formatN. Moreover, patternIn is ISO8601 by default.
  • convDate() is deprecated
  • add(valueToAdd), mul(valueToMultiply), sub(valueToSubstract),div(value) : mathematical operations
  • substr(start, end) : slice strings
• carbone.set and carbone.render have new options
  • currencySource : default currency of source data. Ex 'EUR'
  • currencyTarget : default target currency when the formatter convCurr is used without target
  • currencyRates : rates, based on EUR { EUR : 1, USD : 1.14 }
• Fix memory leaks: one file descriptor remains opened
• Fix crash when template is not correct
• Now option.lang must use the i18n format 'fr-fr' instead of just 'fr'
• Add fallback to basic sorting when timsort crashes
• Bump debug and moment to fix vulnerabilities
• Remove Support of NodeJS 4

v1.1.1

• Release October 11, 2018
• Better Windows support by improving path detection for soffice and python across all operating platforms. Done by Robert Kawecki (@rkaw92).

v1.1.0

• Release February 26, 2018
• Fix: should find tags even if there is a opening bracket { before the tags
• Fix: accept to nest arrays in XML whereas these arrays are not nested in JSON
• Fix: tags were not parsed if formatters were used directly on d or c like this {d:ifEmpty('yeah')} ...
• Fix: keep the first element of the array if the custom iterator is constant
• Fix a lot of strange bugs when using a filter without iterators in arrays (ex. {d.cities[i=0].temperature})
• Optimization: gain x10 when sorting 1 Million of rows
• Add formatter unaccent to remove accent from string
• carbone.set does not overwrite user-defined translations
• Accepts iteration on non-XML. Example: {d[i].brand} , {d[i+1].brand}
• Add new formatters
  • unaccent() to remove accent from string
  • count() to print a counter in loops. Usage: {d[i].name:count()}
  • convCRLF() to convert text, which contains or , into "real" carriage return in odt or docx document
• Formatters which have the property canInjectXML = true can inject XML in documents
• Return an error in render callback when LibreOffice is not detected
• Get the last object of an array using negative values when filtering with i iterator
  • {d.cities[i=-1].temperature} shows the temperature (if the array is not empty) of the last city
  • {d.cities[i=-2].temperature} shows the temperature of the city before the last
  • ...

V1.0.1

• Release October 13, 2017
• Automatically remove XML-incompatible control codes (U+0000 to U+0008 and U+000B to U+000C and U+000E to U+001F) before inserting data in templates
• Fix: ifEqual did not work correctly with boolean values

V1.0.0

• Release June 7, 2017 - First Public Release, First Public Demo on the Web2day event!
• It loads all lang at startup, and it is able to change the lang at runtime
• Avoid unnecessary synchronous code in carbone.set
• Improve documentation
• Default template path is working directory by default
• Return the list of supported format when an unknown options.convertTo is used
• Accept more input type
• Remove deprecated formatters
• carbone.set takes into account changes on factories and startFactory parameters
• Fix: a report without tags, except lang ones, is translated
• Fix: avoid creating LibreOffice zombies when node crashes
• Fix: avoid using LibreOffice if options.convertTo equals input file extension
• Fix: improve tags detection to avoid removing some XML variable like {DSDSD-232D} used in DOCX
• Fix: now compatible with node v4.5.0+, v6+, v8+

v0.13.1

• Release February 22, 2017
• Access properties of the parent object with two (or more) points .. and then access children properties as usual: {d.cities[i, temp=20]..country.history.sport.value}
• Do not crash when there is a javascript error during building process
• Improve error outputs: detect when an unknown formatter is used, and propose a correction

v0.13.0

• Release February 20, 2017
• Access properties of the parent object with two points .. or more. Use case: conditional printing of properties using filters in nested arrays:
  • {d.cities[i, temp=20]..countryName} prints d.countryName only when the temperature of cities equals 20
• Built-in conditional formatters, which starts by if, stop propagation to next formatters if the condition is true
• New formatters:
  • convEnum(d, type): convert enums to human readable values
  • ifEqual(d, value, messageIfTrue, continueOnSuccess): show message if d == value, and stop propagation to next formatters unless continueOnSuccess is true
  • ifContain(d, value, messageIfTrue, continueOnSuccess): show message if d contains value, and stop propagation to next formatters unless continueOnSuccess is true
  • print(d, message): print message
• New function carbone.renderXML(xmlString, data, options, callback) to render XML directly
• Change the lang dynamically in carbone.render and carbone.renderXML with options.lang = 'fr'. The date formatter is automatically propagated on formatters such as convDate
• Replace module zipfile by yauzl: faster, lighter, asynchrone
• XLSX templates are accepted (beta)
• Parse embedded XLSX and DOCX documents
• Add a tool to search a text within a tag in all reports carbone find :formatterName

v0.12.5

• Release December 1, 2016
• Bump moment.js to 2.17.0
• Add some powerful and tested formatters: ifEmpty, arrayJoin, arrayMap, convDate, lowerCase, upperCase, ucFirst, ucWords
• Fix: in formatters convert, format, addDays, parse: if the date is null or undefined these formatters return null or undefined instead of "Invalid Date"

v0.12.4

• Fix: carbone.render crash if options contains formatName without formatOptionsRaw and formatOptions

v0.12.3

• Fix: on OSX, the LibreOffice 5.2 path has changed
• Possibility to add the source file extension in options when using convert(data, convertTo, options, callback) function. It is mandatory for CSV files otherwise LibreOffice does not understand the file type.

  {
    fieldSeparator  : ',',
    textDelimiter   : '"',
    characterSet    : '76',
    sourceExtension : '.csv'
  }

v0.12.2

• Fixed crash when a document cannot be parsed. It returns an error instead of crashing

v0.12.1

• Fixed crash when a document cannot be converted (ods -> doc for example). It returns an error instead of crashing
• Update documentation to install LibreOffice 5.1 on Ubuntu Server
• Fix: crashed when using a nested undefined object inside a template

v0.12.0

• Upgrade Zipfile, improve some tests for Node.js 4.x compatibility

v0.11.3

• Fixed crash when data contains a lot of nested arrays

v0.11.2

• Fix a performance issue when a template isn't using iterators at all, this issue caused the array of data to become really big in some cases even though only a small portion of the data was really used

v0.11.1

• CarboneJS Server can be used on a remote server. CarboneJS send the document by socket instead of writing a file locally (only when a document conversion occurs)
• CarboneJS Server 0.11 is still compatible with CarboneJS v0.10 clients
• Add two benchmarks tests
• Fix a random failure in a test

v0.10.1

• Allow the comparison operator != in your template: {d.cars[engine.power != 3].name}

• Add a newline at the end of the lang file

• Fix: Translation, upper and lower case order does not depend on user's language anymore

• Do not restart LibreOffice if the document is corrupted (restart it only after 10 consecutive attemps)

• Return an error when the document cannot be converted

• Improve stability in general

• Socket connections have been improved:

  • it manages TLS communications
  • it increases the delay between two connection attempts exponentially until it reaches a maximum delay of 20sec

• Fix bugs in tests

• Compatible with Node v0.12

• Update zipfile dependency

• Now, the repetition algorithm can flatten an array of objects!!

  • it accepts to increment multiple arrays in the same time
  • it repeats automatically all tags that are on the same row of all nested arrays

  Example:

Datas:

```
  [
    {
      'site' : {'label':'site_A'},
      'cars' : [
        {
          'name': 'prius',
          'spec': {'weight': 1},
          'wheels':[
            {'brand':'mich'},
            {'brand':'cont'}
          ]
        },
        {
          'name': 'civic',
          'spec': {'weight': 2},
          'wheels':[
            {'brand':'mich'}
          ]
        },
      ],
    },{
      'site' : {'label':'site_B'},
      'cars' : [{
          'name': 'modelS',
          'spec': {'weight': 1},
          'wheels':[
            {'brand':'mich'},
            {'brand':'uni' },
            {'brand':'cont'}
          ]
        }
      ],
    }
  ];
```

Template:

site	car name	weight	brand
{d[i].site.label}	d[i].cars[i].name}	{d[i].cars[i].spec.weight}	{d[i].cars[i].wheels[i].brand}
{d[i+1].site.label}	d[i+1].cars[i+1].name}	{d[i+1].cars[i+1].spec.weight}	{d[i+1].cars[i+1].wheels[i+1].brand}

Result:

site	car name	weight	brand
site_A	prius	1	mich
site_A	prius	1	cont
site_A	civic	2	mich
site_B	modelS	1	mich
site_B	modelS	1	uni
site_B	modelS	1	cont

• Warning: cannot use moxie-zip 0.0.4 because this commit (https://github.com/spocke/moxie-zip/commit/cbc6af14dc2318bbcfcfa82ebaf183c525346ae2) creates wrong *.ods files. The empty directory Bitmap is replaced by a executable file.

v0.10.0

• Add new conversion options for CSV export

  Now it is possible to pass conversion options to LibreOffice. The attribute convertTo can be an object instead of a string:

  var _options = {
    'convertTo' : {
      'formatName' : 'csv',
      'formatOptions' : {
        'fieldSeparator'    : '+',
        'textDelimiter'     : '"',
        'characterSet'      : '0'
      }
    }
  };
  carbone.render(_filePath, data, _options, function(err, result){
    ...
  });

• Add the possibility to pass raw conversion options to LibreOffice. See documentation: https://wiki.openoffice.org/wiki/Documentation/DevGuide/Spreadsheets/Filter_Options

  var _options = {
    'convertTo' : {
      'formatName' : 'csv',
      'formatOptionsRaw' : '124,34,0'
    }
  };
  carbone.render(_filePath, data, _options, function(err, result){
    ...
  });

• Improve LibreOffice detection algorithm for Linux. Now, it supports any version of LibreOffice

• BUG fix: When two nested arrays were incremented in the same time {d.tab[i+1].subtab[i+1]}, it could generate bad XML in certain cases